Warning: Constant WP_TEMP_DIR already defined in /var/www/html/blogs/glenn/wp-config.php on line 94

Warning: Cannot modify header information - headers already sent by (output started at /var/www/html/blogs/glenn/wp-config.php:94) in /var/www/html/blogs/glenn/wp-includes/feed-rss2.php on line 8
Dell Archives - Glenn Berry https://www.sqlskills.com/blogs/glenn/category/dell/ Semi-random musings about SQL Server performance Sat, 11 Jan 2020 20:52:00 +0000 en-US hourly 1 https://wordpress.org/?v=6.9.4 Dell Inspiron 5480 Laptop Review https://www.sqlskills.com/blogs/glenn/dell-inspiron-5480-laptop-review/ https://www.sqlskills.com/blogs/glenn/dell-inspiron-5480-laptop-review/#respond Mon, 02 Sep 2019 21:44:47 +0000 http://3.209.169.194/blogs/glenn/?p=1692 I recently had a chance to work on someone’s new Dell Inspiron 5480 laptop recently, doing some memory and storage upgrades. I also installed and patched a fresh copy of Windows 10 Professional, updated the main BIOS, updated the drive firmware, and also installed all of the latest device drivers. This was purposely meant to […]

The post Dell Inspiron 5480 Laptop Review appeared first on Glenn Berry.

]]>
I recently had a chance to work on someone’s new Dell Inspiron 5480 laptop recently, doing some memory and storage upgrades. I also installed and patched a fresh copy of Windows 10 Professional, updated the main BIOS, updated the drive firmware, and also installed all of the latest device drivers. This was purposely meant to be a rather modest and affordable system that would still be a decent performer in a relatively small and light form factor.

Original Specifications

As originally configured, this machine had an 8th generation Intel Core i7-8565U (4C/8T) “Whiskey Lake” processor with a base clock speed of 1.80 GHz, a max turbo frequency of 4.60 GHz, and an 8 MB L3 cache. It had one 16 GB stick of DDR4-2666 MHz RAM (with one empty slot), which meant the memory was operating in single-channel mode instead of dual channel mode. For storage, it had a 128 GB SK Hynix BC501 M.2 PCIe 3.0 x2 NVMe drive for the operating system, and a 1TB Seagate ST1000LM035-1RK172 5400 rpm SATA hard drive for extra storage.

These general specifications are a pretty big improvement over a mainstream business laptop from a few years ago. It used to be that machines in this price range would have a 2C/4T processor and a single 2.5” SATA drive bay, probably with a conventional hard drive. Having a 4C/8T processor and two drive bays, one being M.2 PCIe 3.0 x4 NVMe, makes a big difference!


Initial Storage Benchmarks

I wanted to confirm that the machine wasn’t DOA from Dell, so I unboxed it and plugged it in. Everything worked fine out of the box, so I collected some quick CrystalDiskMark results on the two OEM drives. First was the 128GB SK Hynix M.2 PCIe NVMe boot drive. As you can see below, this is quite miserable performance for an M.2 PCIe NVMe SSD.

Sequential Read (Q= 32,T= 1) : 1585.788 MB/s
Sequential Write (Q= 32,T= 1) : 287.443 MB/s
Random Read 4KiB (Q=  8,T= 8) : 277.594 MB/s [67772.0 IOPS]
Random Write 4KiB (Q=  8,T= 8) : 255.548 MB/s [62389.6 IOPS]

Keep in mind that this is a M.2 2230 form factor drive that also only supports two PCIe 3.0 lanes, rather than the four PCIe 3.0 lanes that most M.2 PCIe NVMe drives support. This limits its sequential bandwidth to about 1600 MB/second. It is also a very small capacity drive, which further limits its performance because there are fewer NAND cells than in a larger capacity drive. You really should try to stay away from NAND SSDs that are smaller than 500 GB in size.

Just for comparison, here are the same CrystalDiskMark results for the 1TB Seagate ST1000LM035-1RK172 5400 rpm SATA hard drive that came in the system. Even with a 128 MB cache, these are not impressive results for a laptop hard drive. They are probably better than than the hard drives in many existing laptops, but really, using a conventional magnetic hard drive in a laptop is just a miserable experience that I would avoid unless you have no financial alternative.

Sequential Read (Q= 32,T= 1) : 139.031 MB/s
Sequential Write (Q= 32,T= 1) : 133.234 MB/s
Random Read 4KiB (Q=  8,T= 8) : 1.162 MB/s [283.7 IOPS]
Random Write 4KiB (Q=  8,T= 8) : 1.246 MB/s [304.2 IOPS]

Opening Up the Beast

Once I was done with this initial storage testing, it was time to open up the machine to make some improvements. Working with laptops for hardware modifications can be anywhere from easy to fairly difficult. Some “Ultrabook” style machines are virtually impossible to work on, with non-replaceable components or components that are soldered in place. Just getting them opened without breaking or damaging anything can sometimes be pretty challenging. I recommend that you download the service manual for your machine before you even buy it to make sure you understand what you will be dealing with. I also recommend having proper tools from IFixit.com.

I would rate the difficulty of opening up the Dell Inspiron 5480 as medium. There are three captive phillips screws, and seven removeable phillips screws that must be loosened or removed from the base. After that comes the fun part of getting the plastic base to pop off without breaking anything or damaging the plastic surface. IFixit has a number of “prying and opening tools” that are made of carbon fiber or plastic, so you can open things like this without damaging them.


DSC01361

Figure 1: Inside of Dell Inspiron 5480 with OEM Drives


DSC01363

Figure 2: Inside of Dell Inspiron 5480 with Samsung 970 EVO Plus


After opening up the case, I added a 16GB DDR4-2666 SODIMM that I had purchased for $64.99 at Micro Center. Next, I removed both OEM drives, and I installed the 500GB Samsung 970 EVO Plus M.2 2280 PCIe NVMe drive. I had to move a small retaining clip to a different position because the new drive used the longer 2280 form factor (which means it is 80mm long rather than 30mm long). I purposely did not install the 1TB Samsung 860 EVO SATA SSD at this time, because I was going to be installing a fresh copy of Windows 10. In my experience, Windows 10 will sometimes (without asking) install some things on a second drive if it is present during the initial installation, so I always have just a single physical or logical drive present at this point.


DSC01365

Figure 3: Installing Windows 10 (the first time)


As it turned out, I ended up having to install Windows 10 twice because of how Dell had the storage BIOS settings configured from the factory. They had the storage in RAID mode (which uses Intel storage drivers), which seemed to work and perform decently. The only problem with that was that this confused the Samsung Magician software, which claimed that the Samsung 970 EVO Plus was not supported (which prevented me from using Samsung Magician to check for and install firmware updates on the drive). Switching the BIOS to AHCI mode made Windows 10 refuse to boot, so I just reinstalled it from scratch rather than try to fix it. After getting Windows 10 Professional installed, patched, and upgraded to Version 1903, I was ready to continue.

Upgraded Storage Benchmarks

Once everything was installed, patched, and updated, I ran a fresh set of CrystalDiskMark tests on both new drives. First was the new boot drive, which is the 500GB Samsung 970 EVO Plus. As you can see, this was a huge improvement over the original 128GB SK Hynix drive, especially for write performance!

Sequential Read (Q= 32,T= 1) : 3561.355 MB/s
Sequential Write (Q= 32,T= 1) : 3035.333 MB/s
Random Read 4KiB (Q=  8,T= 8) : 710.060 MB/s [173354.5 IOPS]
Random Write 4KiB (Q=  8,T= 8) : 370.695 MB/s [90501.7 IOPS]

Next was the 1TB Samsung 860 EVO SATA SSD that replaced the 1TB Seagate 5400 rpm hard drive. Again, this was a huge improvement over the OEM hard drive. It also shows how much better a good M.2 PCIe NVMe SSD is compared to a SATA AHCI SSD.

Sequential Read (Q= 32,T= 1) : 561.795 MB/s
Sequential Write (Q= 32,T= 1) : 511.736 MB/s
Random Read 4KiB (Q=  8,T= 8) : 398.251 MB/s [97229.2 IOPS]
Random Write 4KiB (Q=  8,T= 8) : 340.293 MB/s [83079.3 IOPS]

Conclusion

For less than $1200.00 (including the upgraded memory and storage), my friend has a pretty capable machine. It has good CPU performance with a relatively modern Intel 4C/8T processor, 32GB of RAM, and 1.5TB of flash storage. It could easily be upgraded to 64GB of RAM, and up to 6TB of flash storage in the future. It also has a modern 802.11ac WiFi radio and a decent amount of USB 3.0 ports and other connectors. This is not a “top of the line machine”, but it will give a good enough user experience for several years.





The post Dell Inspiron 5480 Laptop Review appeared first on Glenn Berry.

]]>
https://www.sqlskills.com/blogs/glenn/dell-inspiron-5480-laptop-review/feed/ 0
SQL Server 2014 Hardware Analysis Case Study https://www.sqlskills.com/blogs/glenn/sql-server-2014-hardware-analysis-case-study/ https://www.sqlskills.com/blogs/glenn/sql-server-2014-hardware-analysis-case-study/#comments Thu, 05 Feb 2015 20:32:59 +0000 http://3.209.169.194/blogs/glenn/?p=932 Imagine that you have been given the go-ahead to upgrade your entire data platform stack from SQL Server 2008 Enterprise Edition to SQL Server 2014 Enterprise Edition. You need to come up with a recommendation for your new database server hardware, looking to maximize performance while controlling your SQL Server 2014 license costs. To help […]

The post SQL Server 2014 Hardware Analysis Case Study appeared first on Glenn Berry.

]]>
Imagine that you have been given the go-ahead to upgrade your entire data platform stack from SQL Server 2008 Enterprise Edition to SQL Server 2014 Enterprise Edition. You need to come up with a recommendation for your new database server hardware, looking to maximize performance while controlling your SQL Server 2014 license costs.

To help you with that effort, here is an example hardware analysis comparing an existing legacy four-socket server (a Dell PowerEdge R815) with four AMD Opteron 6168 processors to a new four-socket server (a Dell PowerEdge R920) with newer 22nm Intel Xeon E7 v2 Ivy Bridge-EX processors.

For a Dell PowerEdge R920, I would be looking at one of these three processors:

1. Xeon E7-8857 v2   (12 cores, 3.0 GHz base clock speed)

2. Xeon E7-8891 v2   (10 cores, 3.2 GHz base clock speed)

3. Xeon E7-8893 v2   (6 cores, 3.4 GHz base clock speed)

These three candidate processors all have higher base clock speeds and lower physical core counts than some other more common choices, such as the fifteen-core Xeon E7-4890 v2.

The closest equivalent AMD-based system I could find in the TPC-E benchmark results (to the legacy system) was an HP ProLiant BL685c G7 Blade Server with four, 2.2GHz AMD Opteron 6174 processors and 512GB of RAM, with an actual raw TPC-E score of 1464.12. The raw TPC-E score is a good way of measuring the overall CPU capacity of a system.

Dividing this score by the number of physical cores in the system gives us a score/core of 30.5, which is a good measure of single-threaded processor performance. Since the legacy system has slower 1.9GHz AMD Opteron 6168 processors (from the same generation and family), we simply need to adjust for the clock speed difference. Taking 1.9GHz divided by 2.2 GHz is 0.8636. Taking the actual 1464.12 score times 0.8636 gives us an estimated TPC-E score of 1264.46 for the legacy system. Dividing that by 48 physical cores gives an estimated score/core of 26.34 for the legacy system.

There is an actual TPC-E result for a four-socket IBM System x3850 X6 with four, 15-core 2.8GHz Intel Xeon E7-4890 v2 processors and 2TB of RAM, with a raw TPC-E score of 5576.27. Dividing this actual score by 60 physical cores gives us an actual score/core of 92.94.

We can adjust this actual result for the three candidate processors listed above to take into account the difference in core counts and base clock speeds to get estimated TPC-E scores for a four-socket system with each of those processors since they are from the same generation and family.

1. Xeon E7-8857 v2               5576.27 original score, times .80 (core count difference), times 1.0714 (clock speed difference), is 4779.53 divided by 48 total physical cores is 99.57 score/core

2. Xeon E7-8891 v2               5576.27 original score, times .66 (core count difference), times 1.1428 (clock speed difference), is 4233.73 divided by 40 total physical cores is 105.84 score/core

3. Xeon E7-8893 v2               5576.27 original score, times .40 (core count difference), times 1.2142 (clock speed difference), is 2708.28 divided by 24 total physical cores is 112.84 score/core

Comparing the legacy system to the actual new four-socket TPC-E result and my estimates for the other three processors, gives us this summary:

Processor                        TPC-E Score        Score/Core         Total Physical Cores     SQL 2014 License Cost (EE)

Opteron 6168                    1264.46                 26.34                     48                             $329,952.00     ($274,464.00 with AMD Core Factor discount)

Opteron 6174                    1464.12                 30.50                     48                             $329,952.00     ($274,464.00 with AMD Core Factor discount)                        

Xeon E7-4890 v2               5576.27                 92.94                     60                             $395,942.00

Xeon E7-8857 v2               4779.53                 99.57                     48                             $329,952.00       

Xeon E7-8891 v2               4233.73                 105.84                   40                             $274,960.00

Xeon E7-8893 v2               2708.28                 112.84                   24                             $164,976.00

This means that we could choose from having from roughly four times better single-threaded processor performance using the Xeon E7-8893 v2 processor or from having roughly four times more processor capacity using the Xeon E7-8857 v2 processor in a new system compared to the legacy system, depending on which processor we choose. The difference in SQL Server 2014 Enterprise Edition license costs between the different processor choices is quite dramatic. For example, going from the twelve-core processor to the faster ten-core processor lowers your SQL Server license costs by about as much as the actual server would cost.

The post SQL Server 2014 Hardware Analysis Case Study appeared first on Glenn Berry.

]]>
https://www.sqlskills.com/blogs/glenn/sql-server-2014-hardware-analysis-case-study/feed/ 7
Bigger Database Servers Get Faster https://www.sqlskills.com/blogs/glenn/bigger-database-servers-get-faster/ https://www.sqlskills.com/blogs/glenn/bigger-database-servers-get-faster/#comments Sat, 01 Mar 2014 21:01:10 +0000 http://3.209.169.194/blogs/glenn/?p=869 For quite some time, I have been talking about how current Intel-based four-socket database servers have had significantly lower single-threaded processor performance than current Intel-based two-socket database servers. This is because the first generation Intel Xeon E7 processors were using the relatively old 32nm Westmere microarchitecture that was introduced in early 2011 for the initial […]

The post Bigger Database Servers Get Faster appeared first on Glenn Berry.

]]>
For quite some time, I have been talking about how current Intel-based four-socket database servers have had significantly lower single-threaded processor performance than current Intel-based two-socket database servers. This is because the first generation Intel Xeon E7 processors were using the relatively old 32nm Westmere microarchitecture that was introduced in early 2011 for the initial Xeon E7 (Westmere-EX) product line.

These E7 processors also use much lower base and turbo clock speeds than current Xeon E5 v2 processors, which also hurts their single-threaded processor performance. They do have higher overall concurrent load capacity due to higher total memory capacity and more total processor cores, but the individual processor cores in most four-socket servers have been much slower than what you find in a modern two-socket server. Simply put, bigger servers are not faster servers. It is like comparing an eighteen wheeler truck to a Tesla Model S.

Now, that old assessment is going to change somewhat, with the release of the 22nm Intel Xeon E7 Processor v2 Family (Ivy Bridge-EX), and new model servers from the major server vendors that have even higher memory capacity, PCI-E 3.0 support, and 12Gbps SAS/SATA support, along with much faster RAID controllers. These processors are a substantial improvement over the previous generation 32nm Intel Xeon E7 processors (Westmere-EX) that have been available since early 2011.

It will still be possible to configure a new two-socket server, such as a Dell PowerEdge R720, with an appropriate 22nm Intel Xeon E5-2600 Processor v2 Family (Ivy Bridge-EP) processor that will have better single-threaded performance than a new four-socket server such as a Dell PowerEdge R920, but the gap will not be nearly as large as it once was.

The actual good news here for a database professional is the fact that you will be able to have a four-socket server that has as much load capacity as a previous generation, eight-socket server, that also performs nearly as well as a current two-socket server, while paying 25% less for your SQL Server 2012/2014 license costs (compared to a previous generation eight-socket server). This is a pretty big gift from Intel!

A more pessimistic view is that your SQL Server 2012/2014 license costs could rise by 50% as you move from an existing server equipped with four, ten-core Xeon E7-4870 processors (with a total of forty physical cores) to a new server with four, fifteen-core Xeon E7-4890 v2 processors (with a total of of sixty physical cores). For reasons known only to Intel, the lower core count SKUs in the Xeon E7-48xx v2 product family are not “frequency optimized”, meaning they do not have higher clock speeds than the high-end, E7-4890 v2 processor. The base and turbo clock speeds of the best lower core-count SKUs in the E7- 48xx v2 family actually drop off pretty quickly as the core counts go down. The shared-L3 cache sizes also drop off very quickly, as does the processor price, as you can see in Table 1.

Processor Physical Cores L3 Cache Base Clock Turbo Clock Price
E7-4890 v2 15 37.5 MB 2.8GHz 3.4GHz $6,619.00
E7-4860 v2 12 30 MB 2.6GHz 3.2GHz $3,838.00
E7-4830 v2 10 20 MB 2.2GHz 2.7GHz $2,059.00
E7-4820 v2 8 16 MB 2.0GHz 2.5GHz $1,446.00
E7-4809 v2 6 12 MB 1.9GHz N/A $1,223.00

Table 1: Selected Intel E7-48xx v2 Processors

 

With the Xeon E4-48xx v2 product family, you are going to want to choose either the E7-4890 v2 or the E7-4860 v2 model processors in most situations, since the lower core count processors are giving up a substantial amount of performance due to their lower clock speeds and smaller L3 cache sizes. If you really want to reduce your core counts to reduce your SQL Server 2012/2014 license costs, you would be better off with the Intel Xeon E5-26xx v2 product family processors that are used in two socket servers. Another alternative is the upcoming Intel Xeon E5-46xx v2 product family processors that are used in four-socket servers.

Either of those choices would be better than one of the lower core count processors in the E7-48xx v2 product family, at least from a pure processor performance perspective.

Intel also has refreshed the E7-88xx v2 product family that is meant for eight-socket and larger servers. For some reason (probably for HPC use), Intel does have “frequency-optimized”, lower core-count models in this product family, as you can see in Table 2.

Processor Physical Cores L3 Cache Base Clock Turbo Clock Price
E7-8890 v2 15 37.5 MB 2.8GHz 3.4GHz $6,841.00
E7-8857 v2 12 30 MB 3.0GHz 3.6GHz $3,838.00
E7-8891 v2 10 37.5 MB 3.2GHz 3.7GHz $6,841.00
E7-8893 v2 6 37.5 MB 3.4GHz 3.7GHz $6,841.00

Table 2: Selected Intel E7-88xx v2 Processors

 

I could see some scenarios where you might want to get an eight-socket server with the six-core E7-8893 v2, so that you could have the same physical core count, while having double the memory capacity and much better single-threaded processor performance than a four-socket server with the twelve-core E7-4860 v2. The hardware cost would be significantly higher, since you would be buying eight processors for $6,841.00 each instead of four processors at $3,838.00 each, but for many organizations, that would not be a major issue.

Some server vendors may offer the Xeon E7-88xx v2 processors in their four-socket server models, since they are pin-compatible, which would give us a lot more flexibility as far as processor selection goes. I really wish Intel had “frequency-optimized” models in their Xeon E7-48xx v2 product family, to make this even easier.

The post Bigger Database Servers Get Faster appeared first on Glenn Berry.

]]>
https://www.sqlskills.com/blogs/glenn/bigger-database-servers-get-faster/feed/ 3
A SQL Server Hardware Tidbit a Day – Day 28 https://www.sqlskills.com/blogs/glenn/a-sql-server-hardware-tidbit-a-day-day-28/ https://www.sqlskills.com/blogs/glenn/a-sql-server-hardware-tidbit-a-day-day-28/#respond Tue, 30 Apr 2013 15:22:18 +0000 http://3.209.169.194/blogs/glenn/?p=751 For Day 28 of this series, we are going to talk about some factors to consider if you are thinking about building a desktop SQL Server 2012 system for development or testing use. I get lots of questions about this subject, and I have been thinking about it some anyway, hence today’s topic. In many […]

The post A SQL Server Hardware Tidbit a Day – Day 28 appeared first on Glenn Berry.

]]>
For Day 28 of this series, we are going to talk about some factors to consider if you are thinking about building a desktop SQL Server 2012 system for development or testing use. I get lots of questions about this subject, and I have been thinking about it some anyway, hence today’s topic.

In many organizations, old retired rack-mounted servers are repurposed as development and test servers. Sometimes, old retired workstations are used for this purpose. Quite often, these old machines are three to five years old (or even older). For example, you will often find old Dell PowerEdge 1850, PowerEdge 6850, and PowerEdge 1950 servers being used for this purpose. These vintage machines are about four to seven years old, and long out of warranty. Their performance and scalability is quite miserable by today’s standards, even compared to a modern desktop.

For example, a Dell PowerEdge 1850, with two Intel Xeon Irwindale 3.0GHz processors and 8GB of RAM has a 32-bit Geekbench score of about 2250. A Dell PowerEdge 6800 with four Xeon 7140M 3.4GHz processors and 64GB of RAM has a 32-bit Geekbench score of 5023. A newer Dell PowerEdge 1950 with two Intel Xeon 5440 processors and 32GB of RAM will have a 32-bit Geekbench score of about 7500. For comparison, my current main workstation has a 22nm Intel Core i7-3770K processor with 32GB of RAM and a 512GB OCZ Vertex 4 SSD. This system has a 32-bit Geekbench score of 12713.

My argument is that in many situations, given a very limited hardware budget, it may make more sense (for development and testing) to build or buy a new desktop system based on a modern platform rather that using relatively ancient “real” server hardware. Your main limiting factors with a new desktop system will be I/O capacity (throughput and IOPS) and memory capacity, but there are some ways around that..  You should be able to build or buy a very capable test system for less than $1500.00, perhaps far less, depending on how you configure it.

Your two main good choices right now are a 22nm Core i7 Ivy Bridge (using a Core i7-3770 or i7-3770K processor) with an Z77 chipset-based motherboard, or a slightly less expensive Core i5-3570 or i5-3570K processor. The Core i7 will have four cores plus hyper-threading, while the Core i5 will have four cores, but no hyper-threading. Either one of these systems will support up to 32GB of RAM, which is how much you should get (since desktop DDR3 RAM is so affordable).

You need to look at the motherboard features and specifications closely to make sure you get what you need without paying too much for unnecessary features. You want to get a motherboard that has as many SATA ports as possible (preferably newer 6Gbps SATA III ports) with hardware RAID support if possible. At the same time, you don’t really need the premium gaming (such as SLI or Crossfire support) and over-clocking features in a top-of-the line motherboard. The entry level motherboards will usually have fewer SATA ports, which is a good reason to go a little higher in the lineup. You can also buy inexpensive PCI-e SATA III expansion cards to add even more SATA ports. You also want to make sure to get a motherboard with four memory slots, since some entry-level motherboards will only have two slots.

Depending on your motherboard vendor, you might run into driver issues with Windows Server 2012. The problem is not that there are no drivers, but the fact that the motherboard vendors sometimes wrap the actual driver installation programs in their own installation programs that do OS version checking that fails with Windows Server 2012 (since they assume you will be using Windows 7 or Windows 8).

You can buy a large, full tower case, with lots of internal 3.5” drive bays. Then you can buy a number of 1TB Western Digital Black 6Gbps hard drives and/or some consumer grade SSDs, depending on your needs and budget. This will let you have a pretty decent amount of I/O capacity for a relatively low cost. Very fast consumer SSDs are now available for less than $1/GB of space, so you can probably find a way to afford one or more of them for your system.

If you can wait until early to mid June, the 22nm Intel Haswell processors will be available. These will require a new motherboard, but will give you about 5-10% better single-threaded CPU performance at the same clock speed as Ivy Bridge, along with a few other benefits.

The post A SQL Server Hardware Tidbit a Day – Day 28 appeared first on Glenn Berry.

]]>
https://www.sqlskills.com/blogs/glenn/a-sql-server-hardware-tidbit-a-day-day-28/feed/ 0
A SQL Server Hardware Tidbit a Day – Day 25 https://www.sqlskills.com/blogs/glenn/a-sql-server-hardware-tidbit-a-day-day-25/ https://www.sqlskills.com/blogs/glenn/a-sql-server-hardware-tidbit-a-day-day-25/#respond Thu, 25 Apr 2013 14:58:28 +0000 http://3.209.169.194/blogs/glenn/?p=747 For Day 24 of this series, I want to talk about the recent history of Dell rack-mounted servers, to help illustrate how processor performance and server capacity has dramatically improved over the past seven years. Back in 2005-2006, you could buy a two-socket Dell PowerEdge 1850, with two hyper-threaded Intel Xeon “Irwindale” 3.2GHz processors and […]

The post A SQL Server Hardware Tidbit a Day – Day 25 appeared first on Glenn Berry.

]]>
For Day 24 of this series, I want to talk about the recent history of Dell rack-mounted servers, to help illustrate how processor performance and server capacity has dramatically improved over the past seven years.

Back in 2005-2006, you could buy a two-socket Dell PowerEdge 1850, with two hyper-threaded Intel Xeon “Irwindale” 3.2GHz processors and 16GB of RAM (with a total of four logical cores). This was fine for an application or web server, but it did not have the CPU horsepower (the 32-bit Geekbench score was about 2200) or memory capacity for a heavy duty database workload.

Around the same time, you could also buy a four-socket Dell PowerEdge 6850, with four dual-core, Intel Xeon 7040 “Paxville” 3.0GHz processors and 64GB of RAM (with a total of 16 logical cores with hyper-threading enabled). This was a much better choice for a database server because of the additional processor, memory, and I/O capacity compared to a PowerEdge 1850. Even so, its Geekbench score was only about 4400, which is pretty pathetic by today’s standards. Back in 2006-2007, it still made perfect sense to buy a four-socket database server for most database server workloads.

By late 2007, you could buy a two-socket Dell PowerEdge 1950, with two, quad-core Intel Xeon E5450 processors and 32GB of RAM (with a total of eight logical cores) and you would actually have a pretty powerful platform for a database server. A system like this would have a 32-bit Geekbench score of about 8000. The biggest weakness of this system was having only two x8 PCI-E 1.0 expansion slots.

By late 2008, you could buy a four-socket Dell PowerEdge R900, with four, six-core Intel Xeon X7460 processors and 256GB of RAM (with a total of of 24 logical cores). This was a very powerful , but costly platform for a database server, with a 32-bit Geekbench score of around 16500. There are still many of these model servers being used for production purposes, and while they sound impressive, that are actually a very bad choice for an upgrade to SQL Server 2012 because of their high physical core counts and low single-threaded performance. The Xeon X7460 was the last generation of Intel SMP processors, before the NUMA-capable Nehalem was introduced.

By early 2009, you could buy a two-socket Dell PowerEdge R710, with two, quad-core Intel Xeon X5570 processors, and 144GB of RAM (with a total of 16 logical cores) and you would have a very powerful database server platform. This system would have a 32-bit Geekbench score of around 15000. This would give you fairly close to the capacity of a four-socket R900, with better single-threaded performance.

By early 2010, you could buy that same Dell PowerEdge R710, with more powerful six-core Intel Xeon X5680 processors (with a total of 24 logical cores), and push the 32-bit Geekbench score to about 22500. This gives you quite a bit more CPU capacity than the PowerEdge R900 that you bought in late 2008. If you are concerned about 144GB of RAM not being enough memory in the R710, you could buy two R710s, and have nearly triple the CPU capacity of a single R900. This assumes that you can split your database workload between two database servers, by moving databases or doing things like vertical or horizontal partitioning of an existing large database.

Finally, by mid-2012, you could buy a 12th generation, Dell PowerEdge R720, with even faster eight-core Intel Xeon E5-2690 processors (with a total of 32 logical cores), which would push the 32-bit Geekbench score to about 29000.  The R720 has 24 memory slots, so you can have 384GB of RAM with 16GB DIMMs or 768 GB of RAM with more expensive 32GB DIMMs. You also get seven PCI-E 3.0 expansion slots, which gives you more potential I/O bandwidth than you can get with a four-socket server (since they are still using the older PCI-E 2.0 standard).

This gap will open up even more in Q3 of 2013, when the 12-core, 22nm Intel Xeon E5-2600 v2 series (Ivy Bridge-EP) processors are released. These will be pin-compatible with the current E5-2600 series, so they will work with current model servers (probably requiring a BIOS update). They should be available very quickly after Intel releases them.

This overall trend has been continuing over the past several years, with Intel introducing new processors in the two socket space roughly a year ahead of introducing a roughly equivalent new processor in the four socket space. This means that you will get much better single-threaded OLTP performance from a two-socket system than from a four-socket system of the same age (as long as your I/O subsystem is up to par).

Given the choice, I would rather have two, two-socket machines instead of one, four-socket machines in almost all cases. The only big exception would be a case where you absolutely need to have far more memory in a single server that you can get in a two socket machine (a Dell PowerEdge R720 can now go up to 768GB if you are willing to pay for 32GB DIMMs), and you are unable to do any re-engineering to split up your load between two servers.

If you want to dive deeper into this subject, you might want to listen to my latest Pluralsight course, which is SQL Server 2012:Evaluating and Sizing Hardware. You can also contact us if you are interested in expert hardware consulting as you get ready to upgrade your database hardware.

The post A SQL Server Hardware Tidbit a Day – Day 25 appeared first on Glenn Berry.

]]>
https://www.sqlskills.com/blogs/glenn/a-sql-server-hardware-tidbit-a-day-day-25/feed/ 0
A SQL Server Hardware Tidbit a Day – Day 23 https://www.sqlskills.com/blogs/glenn/a-sql-server-hardware-tidbit-a-day-day-23/ https://www.sqlskills.com/blogs/glenn/a-sql-server-hardware-tidbit-a-day-day-23/#respond Tue, 23 Apr 2013 13:07:12 +0000 http://3.209.169.194/blogs/glenn/?p=745 For Day 23 of this series, I am going to briefly discuss hardware RAID controllers, also known as disk array controllers. Here is what Wikipedia has to say about RAID controllers: A disk array controller is a device which manages the physical disk drives and presents them to the computer as logical units. It almost […]

The post A SQL Server Hardware Tidbit a Day – Day 23 appeared first on Glenn Berry.

]]>
For Day 23 of this series, I am going to briefly discuss hardware RAID controllers, also known as disk array controllers. Here is what Wikipedia has to say about RAID controllers:

A disk array controller is a device which manages the physical disk drives and presents them to the computer as logical units. It almost always implements hardware RAID, thus it is sometimes referred to as RAID controller. It also often provides additional disk cache.

Figure 1 shows a typical hardware RAID controller.

PERC Series-7 Controllers

Figure 1: Typical Hardware RAID Controller

For database server use (with recent vintage servers), you usually have an embedded hardware RAID controller on the motherboard, that is used for your internal SAS or SATA drives. It is pretty standard practice to have two internal drives in a RAID 1 array, controlled by the embedded RAID controller, that are used to host the operating system and the SQL Server binaries (for standalone SQL Server instances). This gives you a better level of redundancy against losing a single drive and going down.

If you are using Direct Attached Storage (DAS), you will also have one or more (preferably at least two) hardware RAID controller cards that will look similar to what you see in Figure 1. These cards go into an available PCI-E expansion slot in your server, and then are connected by a relatively short cable to an external storage enclosure (such as you see in Figure 2).

PowerVault MD1220

Figure 2: Dell PowerVault MD1220 Direct Attached Storage Array

Each direct attached storage array will have anywhere from 14 to 24 drives. Figure 2 shows a Dell PowerVault MD1220 storage array. The RAID controller(s) are used to build and manage RAID arrays from these available drives, which eventually are presented to Windows as logical drives, usually with drive letters. For example, you could create a RAID 10 array with 16 drives and another RAID 10 array with eight drives from a single 24 drive direct attached storage array. These two RAID arrays would be presented to Windows, and show up as say the L: drive and the R: drive.

Enterprise level RAID controllers usually have some cache memory on the card itself. This cache memory can be used to cache reads or to cache writes, or split between both. For SQL Server OLTP workloads, it is a standard best practice to devote your cache memory entirely to write caching. You can also choose between write-back and write-through cache policies for your controller cache. Write-back caching provides better performance, but there is a slight risk of having data in the cache that has not been written to the disk if the server fails. That is why it is very important to have a battery-backed cache if you decide to use write-back caching.

Most enterprise-level RAID controllers will fall-back from write-back caching to write-though caching (which is safer, but slower) if the battery for the cache is not present and charged. Some newer, high-end RAID controllers are also able to use a feature developed by LSI called CacheCade that lets you use a number of SSDs as a cache in front of conventional SAS drives. This gives you much of the performance benefit of SSD storage without having to spend the money to have 100% SSD storage.

The post A SQL Server Hardware Tidbit a Day – Day 23 appeared first on Glenn Berry.

]]>
https://www.sqlskills.com/blogs/glenn/a-sql-server-hardware-tidbit-a-day-day-23/feed/ 0
A SQL Server Hardware Tidbit a Day – Day 22 https://www.sqlskills.com/blogs/glenn/a-sql-server-hardware-tidbit-a-day-day-22/ https://www.sqlskills.com/blogs/glenn/a-sql-server-hardware-tidbit-a-day-day-22/#comments Mon, 22 Apr 2013 22:21:33 +0000 http://3.209.169.194/blogs/glenn/?p=744 I have written previously about Building a Workstation for SQL Server 2012 Development and Testing and Building a Larger Capacity SQL Server 2012 Workstation. In those posts, I talked about the fact that it becomes increasingly expensive to build a system from parts as you move beyond mainstream desktop systems based on an Intel Core […]

The post A SQL Server Hardware Tidbit a Day – Day 22 appeared first on Glenn Berry.

]]>
I have written previously about Building a Workstation for SQL Server 2012 Development and Testing and Building a Larger Capacity SQL Server 2012 Workstation. In those posts, I talked about the fact that it becomes increasingly expensive to build a system from parts as you move beyond mainstream desktop systems based on an Intel Core i7-3770 or i7-3770K processor.

If you actually need more than 32GB of RAM or more than four physical processor cores, you will have to move up to a single-socket, Sandy Bridge-E, Socket 2011 system, which maxes out at 64GB of RAM and six physical processor cores, at a considerably higher cost. The next step up is moving to a single-socket, Sandy Bridge-EP system with one Intel Xeon E5-1600 series or E5-2600 series processor with up to eight physical processor cores and 128GB of RAM.

Finally, you can buy the components for a dual-socket, Socket 2011 system with two Intel Xeon E5-2600 series Sandy Bridge-EP processors to build a much more powerful and expensive system with up to 16 physical processor cores and 256GB of RAM. By the time you do this, you may end up spending so much money that you would be much better off buying a refurbished, actual server from someplace like the Dell Outlet.

Depending on what you are planning on doing with your server, you can choose a Tower server, a Rack server, or even a Blade server. Depending on your infrastructure, you are probably more likely to be looking at a tower server or a rack server.

Dell Outlet PowerEdge Tower Servers

Dell Outlet PowerEdge Rack Servers

One tactic I like to use when I am searching for a server on the Dell Outlet is to focus on the installed processors in the server. You are very unlikely to ever upgrade the processors in a server after you buy it since processor upgrade kits are usually quite expensive. That means you should focus on the processors and not worry as much about the amount of RAM. Getting more RAM later is easy and inexpensive.

Storage is a little trickier. Sometimes you can find outlet servers that have a lot of big, fast internal drives that are bargains. You may also be better off to think about buying a number Intel DC S3700 SSDs along with compatible 2.5” drive carriers from someplace like Amazon (since Dell does not seem to like selling empty drive carriers very much). You should be on the lookout for systems that have decent RAID controllers in them, since they are relatively expensive to buy later.

As you get ready to buy a server from the Dell Outlet, you want to make sure you have the necessary means of payment ready to go as soon as you find the system you want, since their available inventory is constantly changing. If your organization has a lot of bureaucratic overhead, it may take too long for you to get the necessary approvals before that system you were interested in is no longer available.

The post A SQL Server Hardware Tidbit a Day – Day 22 appeared first on Glenn Berry.

]]>
https://www.sqlskills.com/blogs/glenn/a-sql-server-hardware-tidbit-a-day-day-22/feed/ 2
Two Database Server Models Compared https://www.sqlskills.com/blogs/glenn/two-database-server-models-compared/ https://www.sqlskills.com/blogs/glenn/two-database-server-models-compared/#comments Wed, 24 Oct 2012 14:49:33 +0000 /blogs/glenn/post/Two-Database-Server-Models-Compared.aspx If you have read my SQL Server Hardware book, or ever heard me do one of my hardware presentations, you have probably heard my exhortations to “Never let anyone talk you into reusing old hardware for a new version of SQL Server”.  This is especially true with SQL Server 2012 Enterprise Edition and its core-based […]

The post Two Database Server Models Compared appeared first on Glenn Berry.

]]>
If you have read my SQL Server Hardware book, or ever heard me do one of my hardware presentations, you have probably heard my exhortations to “Never let anyone talk you into reusing old hardware for a new version of SQL Server”.  This is especially true with SQL Server 2012 Enterprise Edition and its core-based licensing model. In order to illustrate this point, I want to compare a couple of different, common server models from Dell, from a SQL Server 2012 Enterprise Edition perspective. The reason that I focus on Enterprise Edition is because of the 64GB RAM limit for Standard Edition and because of all of the compelling features in Enterprise Edition that are so valuable for mission critical database servers.

The Contenders

The “old” model server that I want to use for this comparison is a Dell PowerEdge R900 server, which is a 10th generation, four-socket, 4U, rack-mount server that used the Intel Xeon 7200 series, Intel Xeon 7300 series and the Intel Xeon 7400 series processors, culminating in the Intel Xeon X7460 processor that was released in Q3 of 2008. This processor was the hot ticket in late 2008 through early 2010 (when the Nehalem-based Intel Xeon 7500 series was introduced in Q1 of 2010). I can remember unsuccessfully begging my CTO at NewsGator to let me get a new R900 server during that time frame, but I never won that battle!  I often see customers that are still using this model server in production, so it is not ancient by any means.

The current model server I am going to use for the comparison is the Dell PowerEdge R720xd server, which is a 12th generation, two-socket, 2U, rack-mount server that uses the Intel Xeon E5-2600 series processors. The top of the line model for the Intel Xeon E5-2600 series is the Intel Xeon E5-2690, which is the current champion for single-threaded processor performance according to recent TPC-E OLTP benchmark results.

Comparing the Processors

The 45nm Intel Xeon X7460 “Dunnington” is a six-core processor that has a clock speed of 2.66GHz. It does not have Intel Hyper-Threading or Turbo Boost Technology, and it does not support non-uniform memory access (NUMA). In fact, it was the last generation processor in the 7000 sequence to have a symmetric multiprocessing (SMP) architecture. SMP machines had increasing problems with memory contention causing performance and scalability issues as the number of sockets increased in the server, especially once you went above four sockets.

The 32nm Intel Xeon E5-2690 “Sandy Bridge-EP” is an eight-core processor that has a base clock speed of 2.9GHz. It can use Turbo Boost to increase the speed of individual processor cores up to 3.8GHz, and it has Hyper-Threading so that you will have 16 logical processors available to the operating system. The Xeon E5-2690 does have NUMA support, and the on-board memory controller has four memory channels. The E5-2690 has support for the newer PCI-E 3.0 standard which gives twice the total bandwidth of the older PCI-E 2.0 standard.  This is a much better processor than the old Xeon X7460.

Comparing the Server Capacities

The Dell PowerEdge R900 server could have four, six-core Intel Xeon X7460 processors for a total of 24 physical cores in the system. The R900 has 32 DIMM memory slots that can each hold 8GB FB-DIMMs for a total of 256GB of RAM in the system. There are seven PCI-E 2.0 slots in this server, with four x8 slots and three x4 slots. There are also eight 2.5” internal drive bays that support 6Gbps SAS drives.

The Dell PowerEdge R720xd server could have two, eight-core Intel Xeon E5-2690 processors for a total of 16 physical cores in the system. With hyper-threading enabled, you would have 32 logical cores. The R720xd has 24 DIMM memory slots that each hold 32GB DDR3 ECC DIMMs for a total of 768GB of RAM. Realistically, it does not make economic sense to use 32GB DIMMs since they still cost about $1500 each! That means you would likely choose 16GB DIMMs for a total capacity of 384GB of RAM in the system. There are six PCI-E 3.0 slots in this server, with two x16 slots and four x8 slots. There are also (26) 2.5” internal drive bays that support 6Gbps SAS drives. The R720xd supports 50% more RAM (using affordable 16GB DIMMs), triple the PCI-E slot bandwidth, and more than triple the internal drive bay capacity compared to the R900.

Comparing the TPC-E Performance

There is actually a TPC-E submission for a Dell PowerEdge R900 server with four Xeon X7460 processors from August 19, 2008, with a score of 671.35. Dividing 671.35 by 24 physical cores gives us a score of 27.97 per physical core, which is quite low by modern standards.

There are no TPC-E submissions for a Dell PowerEdge R720xd server, so I picked one for a Fujitsu PRIMERGY RX300 S7 with two Xeon E5-2690 processors from July 5, 2012, with a score of 1871.81. Dividing 1871.81 by 16 physical cores gives us a score of 116.99 per physical core, which is much better than the older R900 system!

Comparing the SQL Server 2012 Enterprise Edition Licensing Costs

The Dell PowerEdge R900 server with four Xeon X7460 processors has 24 physical processor cores that will each require a SQL Server 2012 Enterprise Edition core license that costs $6874.00. This will cost you $164,976.00 for the SQL Server 2012 license costs.

The Dell PowerEdge R720xd server with two Xeon E5-2690 processors has 16 physical processor cores that will each require a SQL Server 2012 Enterprise Edition core license that costs $6874.00. This will cost you $109,984.00 for the SQL Server 2012 license costs.  The savings in license costs compared to the R900 would pay for the server and leave about $35K available for other uses.

Comparing the SQL Server 2008 R2 Enterprise Edition Licensing Costs

The Dell PowerEdge R900 server with four Xeon X7460 processors has 4 physical processor sockets that will each require a SQL Server 2008 R2 Enterprise Edition processor license that costs $27,999.00. This will cost you $111,996.00 for the SQL Server 2008 R2 license costs.

The Dell PowerEdge R720xd server with two Xeon E5-2690 processors has 2 physical processor sockets that will each require a SQL Server 2008 R2 Enterprise Edition processor license that costs $27,999.00. This will cost you $55,998.00 for the SQL Server 2008 R2 license costs. The savings in license costs compared to the R900 would pay for the server and leave about $35K available for other uses.  Even if you are going to use SQL Server 2008 R2 Enterprise Edition instead of SQL Server 2012 Enterprise Edition, you will still save money and have a faster server with more overall capacity by buying the new server.

Conclusions

Reusing your existing Dell PowerEdge R900 server for SQL Server 2012 Enterprise Edition would cost you $54,992.00 in additional SQL Server 2012 license costs to have about 36% of the TPC-E performance, compared to buying a new Dell PowerEdge R720xd server with two Xeon E5-2690 processors. This new server would cost roughly $15-20K depending on how you configured it and how much of a discount you got from your sales representative. This new server could have 50% more RAM (with 16GB DIMMs), triple the PCI-E slot bandwidth, and more than triple the internal drive bay capacity compared to the R900.  The R720xd server would also use less power, and only take up 2U of rack space.  Your net savings from buying the new server would be in the $35-40K range. If this is not good enough, I have an alternate configuration that would save even more money!

Instead of using the eight-core Xeon E5-2690 processors, you could choose the four-core Xeon E5-2643 processors for your new Dell R720xd server. This would reduce your hardware cost for the new server by about $2000.00 (since the processors are less expensive), and it would reduce your SQL Server 2012 licensing costs by 50%, since you would only have to have eight processor core licenses for a total of $54,992.00. This would give you a net savings of about $91-96K if you bought the new server. Since the E5-2643 has a base clock speed of 3.3GHz (with Turbo Boost to 3.5Ghz), you would actually see better single-threaded performance than with the E5-2690. You would give up some capacity and scalability compared to the E5-2690, but it would still be a big improvement over the existing R900 server. I would estimate that a two socket-server with two Xeon E5-2643 processors would have a TPC-E score of about 1029, which is still significantly better than the old R900 server.

The post Two Database Server Models Compared appeared first on Glenn Berry.

]]>
https://www.sqlskills.com/blogs/glenn/two-database-server-models-compared/feed/ 2
Dell 12th Generation Rack Mounted Server Comparison https://www.sqlskills.com/blogs/glenn/dell-12th-generation-rack-mounted-server-comparison/ https://www.sqlskills.com/blogs/glenn/dell-12th-generation-rack-mounted-server-comparison/#respond Thu, 27 Sep 2012 22:05:13 +0000 /blogs/glenn/post/Dell-12th-Generation-Rack-Mounted-Server-Comparison.aspx I think it is useful to examine the various 12th generation, rack-mounted server models available from Dell, comparing them from a SQL Server perspective. There are currently seven models in this line, ranging from an entry-level, single-socket server to a four-socket server. All of these servers use the Intel Xeon E5 processor (Sandy Bridge), but […]

The post Dell 12th Generation Rack Mounted Server Comparison appeared first on Glenn Berry.

]]>
I think it is useful to examine the various 12th generation, rack-mounted server models available from Dell, comparing them from a SQL Server perspective. There are currently seven models in this line, ranging from an entry-level, single-socket server to a four-socket server. All of these servers use the Intel Xeon E5 processor (Sandy Bridge), but different models use different series of that processor family, which is a very important  detail to pay attention to for SQL Server 2012 usage. You could easily compare the available server models from another vendor in the same way, and I will probably do that in a future post.

Dell PowerEdge R320
This model server has a 1U form factor, one processor socket, and uses the Intel Xeon E5-2400 series processor. It has six memory slots (96GB total RAM with 16GB DIMMs), has eight 2.5 inch drive bays, and has one PCI-E 3.0 x8 and one PCI-E 3.0 x16 expansion slots. It has a total of four, six, or eight physical cores for SQL Server 2012 core licensing purposes. It has a total of eight, twelve, or sixteen logical cores with Intel hyper-threading enabled. The R320 is an interesting option for some smaller workloads, since it uses the Xeon E5-2400 series Sandy Bridge-EN processor (that is usually used in two-socket systems) instead of the Xeon E3-1200 Sandy Bridge or Xeon E3-1200 v2 series Ivy Bridge processor that is used in most new single-socket servers. This lets you use up to 96GB of RAM instead of being limited to 32GB of RAM, and it lets you have up to eight physical processor cores instead of being limited to four physical processor cores. The downside of this is being limited to slower processor clock speeds with the E5-2400 series compared to the E3-1200 series processors, which means you will see slower single-threaded performance. The R320 might be a good choice for a DW type of workload, where the extra processor cores and higher memory capacity would be more useful. A single-socket server with an Intel E3-1200 v2 series processor would be better for a small OLTP workload. Keep in mind that SQL Server 2012 Standard Edition is limited to 64GB of RAM.

Dell PowerEdge R420
This model server has a 1U form factor, two processor sockets, and uses Intel Xeon E5-2400 series processors. It also has twelve memory slots (192GB total RAM with 16GB DIMMs), has eight 2.5 inch drive bays, and has two PCI-E 3.0 x16 expansion slots. It has a total of eight, twelve, or sixteen physical cores for SQL Server 2012 core licensing purposes. It has a total of sixteen, 24, or 32 logical cores with Intel hyper-threading enabled. This model is a bad choice for SQL Server 2012. The Xeon E5-2400 series  Sandy Bridge-EN processor is very limited  compared to the Xeon E5-2600 series Sandy Bridge-EP processor. It has slower clock speeds, less memory bandwidth, and less memory capacity. Since you pay the same amount for each SQL Server 2012 core license regardless of what type of physical core is in the processor, the E5-2400 series is a bad choice compared to the E5-2600 series. Another problem with this model server is the fact that it only has two PCI-E expansion slots and eight internal drive bays, which limits your overall I/O capacity and performance.

 

Dell PowerEdge R520
This model server has a 2U form factor, two processor sockets, and uses Intel Xeon E5-2400 series processors. It also has twelve memory slots (192GB total RAM with 16GB DIMMs), has eight 3.5 inch drive bays, and has three PCI-E 3.0 x8 and one PCI-E 3.0 x16 expansion slots. It has a total of eight, twelve, or sixteen physical cores for SQL Server 2012 core licensing purposes. It has a total of sixteen, 24, or 32 logical cores with Intel hyper-threading enabled. This model is also a bad choice for SQL Server 2012 since it uses the same Intel Xeon E5-2400 series processor as the R420 . It does have four PCI-E expansion slots, which is a little better for I/O capacity and performance. Still, I would steer clear of both the R420 and R520 models for SQL Server 2012 usage.

 

Dell PowerEdge R620
This model server has a 1U form factor, two processor sockets, and uses Intel Xeon E5-2600 series processors. It has 24 memory slots (384GB total RAM with 16GB DIMMs), has ten 2.5 inch drive bays, and has one PCI-E 3.0 x8 and two PCI-E 3.0 x16 expansion slots. It also has a total of eight, twelve, or sixteen physical cores for SQL Server 2012 core licensing purposes. It has a total of sixteen, 24, or 32 logical cores with Intel hyper-threading enabled. The R620 is a much better choice for SQL Server 2012 than either the R420 or R520 since it uses the Intel Xeon E5-2600 series Sandy Bridge-EP processor. That processor series gives you higher clock speeds, higher memory bandwidth, and higher memory capacity compared to the Xeon E5-2400 series Sandy Bridge-EN processor. The R620 is limited to three PCI-E expansion slots, but it does have ten internal drive bays. Overall, it is a good model for use as an entry level two-socket database server, especially if you want a 1U form factor.

 

Dell PowerEdge R720
This model server has a 2U form factor, two processor sockets, and uses Intel Xeon E5-2600 series processors. It has 24 memory slots (384GB total RAM with 16GB DIMMs), has sixteen 2.5 inch drive bays, and has six PCI-E 3.0 x8 and one PCI-E 3.0 x16 expansion slots. It has a total of eight, twelve, or sixteen physical cores for SQL Server 2012 core licensing purposes. Total of sixteen, 24, or 32 logical cores with Intel hyper-threading enabled. The R720 is one of my favorite models in the Dell 12th generation line. It uses the same Intel Xeon E5-2600 series processor as the R620, but it has seven PCI-E expansion slots and sixteen internal drive bays, which combine to give you a lot of potential I/O capacity and performance. It does cost a little bit more than the R620, and it is in a 2U vertical size, so there are some scenarios where I would prefer an R620. An example scenario would be an OLTP workload where I knew that I would have external SAN storage with very good random I/O performance, and I wanted to be able to use 1U database servers instead of 2U database servers.

 

Dell PowerEdge R720xd
This model server has a 2U form factor, two processor sockets, uses Intel Xeon E5-2600 series processors, has 24 memory slots (384GB total RAM with 16GB DIMMs), has 26 2.5 inch drive bays, and has four PCI-E 3.0 x8 and two PCI-E 3.0 x16 expansion slots. Has a total of eight, twelve, or sixteen physical cores for SQL Server 2012 core licensing purposes. Total of sixteen, 24, or 32 logical cores with Intel hyper-threading enabled. The R720xd is similar to the R720, except that it has 26 internal drive bays and only six PCI-E expansion slots. This model could be a good choice if you can run your I/O workload on 26 internal drive bays, some or all of which could be solid state drives (SSDs). This could let you avoid the expense of an external direct-attached storage (DAS) enclosure or a storage area network (SAN).

 

Dell PowerEdge R820
This model server has a 2U form factor, four processor sockets, and uses Intel Xeon E5-4600 series processors. It has 48 memory slots (768GB total RAM with 16GB DIMMs), has sixteen 2.5 inch drive bays, and has five PCI-E 3.0 x8 and two PCI-E 3.0 x16 expansion slots. It also has a total of sixteen, 24, or 32 physical cores for SQL Server 2012 core licensing purposes. Total of 32, 48, or 64 logical cores with Intel hyper-threading enabled. The R820 has four processor sockets in a 2U vertical size. It uses the Intel Xeon E5-4600 series processor, which has lower clock speeds than the Xeon E5-2600 series. There is also some non-uniform memory access (NUMA) scaling loss as you move from a two-socket to a four-socket server, i.e. a four socket server does not have twice the scalability as a two socket server with the exact same processor. The R820 does have sixteen internal drive bays and seven PCI-E expansion slots, so it has good I/O capacity and performance potential. It also has twice the total RAM capacity compared to an R620, R720, or R720xd. In spite of all these factors, I would tend to prefer two R720xd servers instead of one R820 server, assuming you can split your workload between two servers. You would have faster, less expensive processors, over three times as many internal drive bays, and nearly twice as many PCI-E expansion slots, while paying the same SQL Server 2012 license costs. I really like the R720xd, with its 26 internal drive bays. I suspect that a very high percentage of SQL Server workloads would run extremely well on an R720xd. If 26 internal drives did not give you enough I/O performance and capacity, you could always add some internal solid state storage cards or use some form of external storage.

As a database professional, I would be actively lobbying against using the R420 or R520 models, since they have the entry-level Intel Xeon E5-2400 series processors, which have lower clock speeds and less memory bandwidth compared to the Intel Xeon E5-2600 series processors that are used in the R620, R720, and R720xd. They also have half of the total memory capacity and far fewer PCI-E slots compared to the higher end models. They are a little less expensive, but the hardware cost delta is very small compared to the SQL Server 2012 license costs. Remember, you are paying for SQL Server 2012 core licenses based on physical core counts, so you want to get the best package you can as far as the rest of the server goes. One nice fact is that the Intel Xeon E5 processor family is available in four-core, six-core, and eight core models, with specific four-core models having higher base clock speeds than the “top-of-the line” eight-core model processor. If you wanted to minimize your SQL Server 2012 core-based licensing costs and were willing to give up some scalability and capacity, you could pick one of these faster base clock speed four-core model processors for your server and actually see very good single-threaded performance.

A good example of a fast, quad-core Intel Xeon E5 processor is the Intel Xeon E5-2643 that runs at a base clock speed of 3.3GHz, with a Turbo Boost speed of 3.5GHz. This processor would give very good OLTP performance at 50% the SQL Server 2012 Core license cost of an eight-core Intel Xeon E5-2690 that runs at a base clock speed of 2.9GHz, with a Turbo Boost speed of 3.8GHz.

The post Dell 12th Generation Rack Mounted Server Comparison appeared first on Glenn Berry.

]]>
https://www.sqlskills.com/blogs/glenn/dell-12th-generation-rack-mounted-server-comparison/feed/ 0