Leaked Information About AMD Opteron 6200 Series Processors

CPU World has a post about some leaked info regarding the upcoming Opteron 6200 series “Bulldozer” processor line. Here are the specifications of the four models that the story describes:

Opteron 6220       8 cores, 3.0GHz base clock speed, Turbo CORE to 3.5GHz

Opteron 6234      12 cores, 2.4GHz base clock speed, Turbo CORE to 2.9GHz

Opteron 6272      16 cores, 2.1GHz base clock speed, Turbo CORE to 2.6GHz, 16MB L3 cache

Opteron 6276      16 cores, 2.3GHz base clock speed, Turbo CORE to 2.8Ghz

AMD explains how Turbo CORE works here.  I really hope that the Bulldozer is released soon, and that it performs well. Otherwise, Intel will continue to completely dominate the high end portion of the server market. Viable competition is good for the DBA community…

Posted in Computer Hardware, Processors | Tagged | Leave a comment

Building or Buying a High Performance, Low Cost Desktop Machine for SQL Server Development and Testing

There was a lengthy discussion on Twitter this morning, where one of my friends, Grant Fritchey (Twitter) was asking the community for suggestions about a good hardware setup that would enable him to run two to three concurrent virtual machines, all running SQL Server. My initial thoughts ran to buying or building either an Intel Socket 1366 or Socket 1155 platform based desktop system that can be used as a development or test server. With a machine based on either one of the platforms, you will have more CPU capacity than many production database servers. You will be limited somewhat by RAM capacity, but your more likely bottleneck will be I/O capacity, compared to a real rack mounted server connected to a good I/O subsystem.

Intel Socket 1366 Platform

This platform is a little older, but it does have six DDR3 memory slots. It supports a maximum of 24GB of RAM, which should be sufficient for a number of concurrent virtual machines. The newer 1366 motherboards usually have two 6Gbps SATA III ports (that unfortunately use a slower Marvell controller), and a couple of USB 3.0 ports. There are two likely processors that I would choose for a 1366 motherboard, the Core i7-970 or the Core i7-960. Either one of these processors would work well in one of the newer Socket 1366 motherboards like the ASUS Sabertooth X58 LGA 1366 Intel X58 SATA 6Gb/s USB 3.0 ATX Intel Motherboard ($184.99 at NewEgg).

The more expensive one is the 3.2GHz 32nm Intel Core i7-970 “Gulftown” ($579.99 at NewEgg).  This CPU has six cores, plus hyper-threading, so it has 12 logical cores. It is relatively affordable for a Gulftown CPU (compared to around $1000 for a Core i7-990X), but it does not offer twice the capacity of the Core i7-960, even though it costs twice as much. It is the desktop equivalent of an Intel Xeon W3670.

The more affordable one is the 3.2GHz 45nm Intel Core i7-960 “Bloomfield” ($284.99 at NewEgg).  This CPU has four cores, plus hyper-threading, so it has 8 logical cores. I would argue that it will give you much more bang for the buck compared to the Core i7-970. It is the desktop equivalent of an Intel Xeon W3570.

Intel Socket 1155 Platform

This is the newer, initial desktop Sandy Bridge platform, with four DDR3 memory slots. It supports a maximum of 32GB of RAM (if you can find and spring for 8GB sticks of DDR3 RAM), but it is more likely you will be using more economical 4GB sticks right now, so you will be limited to 16GB of RAM. This should be plenty of RAM for three or four concurrent virtual machines. The better Socket 1155 motherboards will have two native 6Gbps SATA III ports, plus two more Marvell 6Gbps SATA III ports. They also typically have six 3Gbps SATA II ports, usually with hardware RAID support. A good example is the ASUS P8Z68-V PRO LGA 1155 Intel Z68 HDMI SATA 6Gb/s USB 3.0 ATX Intel Motherboard ($209.99 at NewEgg). The best processor for this platform right now is the Core i7-2600K.

The 3.4GHz 32nm Intel Core i7-2600K “Sandy Bridge” ($314.99 at NewEgg), is the best Sandy Bridge desktop CPU available right now. It has four cores, plus hyper-threading for a total of eight logical cores. It uses the newer Turbo Boost 2.0 to go up to 3.8GHz on individual cores. It is the desktop equivalent of the Intel Xeon E3-1275. A Core i7-2600K processor will have roughly equivalent CPU performance to a Core i7-970, for about half the cost. It will also use less electrical power, and run cooler. The Sandy Bridge processors have pretty decent integrated graphics that are more than sufficient for a desktop “server” machine. Depending on which motherboard chipset you choose (either H67 or Z68 based), you can choose to use the integrated graphics instead of a discrete graphics card. This saves electrical consumption and reduces your hardware cost, but the integrated graphics will use a little bit of your available RAM.

Currently, the DDR3 RAM used by both the 1366 and 1155 platforms is very affordable. For example, you can get a 16GB kit for $139.99. The next big choice is the type and quantity of storage that you decide to get, which depends on your needs and budget. Solid State Drives (SSD) are becoming more affordable, but they are still pretty expensive compared to conventional rotating disk drives. On the other hand, they offer much better I/O performance, which is very important for virtual machines. The new generation 6Gbps SATA III SSDs offer much better throughput performance than the older 3Gbps SATA II SSDs, especially when they are plugged into a 6Gbps SATA III port, but any SSD is going to offer excellent performance compared to a magnetic drive. The larger capacity, (and more expensive) SSDs perform better than the small capacity models in the same line because they use more flash memory chips with more internal I/O channels.

If you have money to burn, you can get a 480GB OCZ Vertex 3 SSD for $1199.00 at NewEgg. A slightly more affordable alternative would be a 240GB OCZ Vertex 3 Max IOPS SSD for $539.99 at NewEgg. Going down in price, you can get a somewhat older 128GB Crucial RealSSD C300 SSD for $231.99 at NewEgg. Depending on your space and performance needs (and your budget), you might want to use a mixture of different storage types, starting with a relatively small and affordable SSD boot drive combined with several large traditional hard drives. I really like the larger Western Digital Black drives, such as the 2TB WD2002FAEX for $149.99 at NewEgg.  When you look at these big WD Black drives, make sure to get the 6Gbps models with the 64MB cache instead of the older 3Gbps models with a 32MB cache.

If you are going to build a system from parts (which is really not very hard), you will also need a case, power supply, and an optical drive. You also might need an inexpensive discrete graphics card. If you don’t want to mess around with building a system yourself, you can either buy a “white box” system that uses similar components,or you could buy a system from a larger vendor, and possibly add more RAM and drives as needed. One way or another, you can have a very powerful system for anywhere from $1000 to $1500.

Posted in Computer Hardware, Processors, SQL Server 2008, SQL Server 2008 R2, SQL Server Denali | Tagged | 15 Comments

Microsoft SQL Server 2008 SP2 Cumulative Update 5 Released

Microsoft has released Cumulative Update 5 for SQL Server 2008 Service Pack 2. This CU has ten fixes listed in the knowledge base article, and it is Build 10.0.4316. This Cumulative Update is only for SQL Server 2008 SP2 (not for SQL Server 2008 R2).

Microsoft has also released Cumulative Update 15 for SQL Server 2008 Service Pack 1, with only four fixes listed in the KB article. This is Build 10.0.2847. Of course, it should not be a big surprise that I think you should be on the SQL Server 2008 SP2 branch by now. At some point, Microsoft will “retire” the SQL Server 2008 SP1 branch, making it an “unsupported service pack”.

Here is an updated SQL Server 2008 Build Chart that makes it a little easier to understand:

-- SQL Server 2008 RTM is considered an "unsupported service pack" as of April 13, 2010
-- SQL Server 2008 RTM Builds   SQL Server 2008 SP1 Builds     SQL Server 2008 SP2 Builds
-- Build       Description      Build       Description         Build     Description
-- 1600        Gold RTM
-- 1763        RTM CU1
-- 1779        RTM CU2
-- 1787        RTM CU3    -->    2531        SP1 RTM
-- 1798        RTM CU4    -->    2710        SP1 CU1
-- 1806        RTM CU5    -->    2714        SP1 CU2 
-- 1812        RTM CU6    -->    2723        SP1 CU3
-- 1818        RTM CU7    -->    2734        SP1 CU4
-- 1823        RTM CU8    -->    2746        SP1 CU5
-- 1828        RTM CU9    -->    2757        SP1 CU6
-- 1835        RTM CU10   -->    2766        SP1 CU7
-- RTM Branch Retired     -->    2775        SP1 CU8      --> 4000       SP2 RTM
--                                2789       SP1 CU9
--                                2799       SP1 CU10    
--                                2804       SP1 CU11    -->  4266       SP2 CU1        
--                                2808       SP1 CU12    -->  4272       SP2 CU2    
--                                2816       SP1 CU13    -->  4279       SP2 CU3    
--                                2821       SP1 CU14    -->  4285       SP2 CU4    
--                                2847       SP1 CU15    -->  4316       SP2 CU5            
Posted in Microsoft, SQL Server 2008 | Tagged | Leave a comment

SQL Server 2008 R2 Service Pack 1 Released

Microsoft has finally (after releasing the CTP version way back in April) released the final, gold build of SQL Server 2008 R2 SP1. This service pack includes the fixes from SQL Server 2008 R2 RTM Cumulative Updates 1 through 6, plus some new features that are not included in those Cumulative Updates. A TechNet Wiki with Release Notes for SQL Server 2008 R2 SP1 is available here.

If you have already installed SQL Server 2008 R2 RTM CU7 or CU8, you should hold off on installing SQL Server 2008 R2 SP1 until SQL Server 2008 R2 SP1 CU1 is released (probably in about six weeks), since that will probably be synchronized with SQL Server 2008 R2 RTM CU9, if history is any judge.   If you are still on build 10.50.1600 or an earlier CU than CU7, you can go ahead and install Service Pack 1. The little build chart below might make it more clear.

 

SQL Server 2008 R2 Builds                     SQL Server 2008 R2 SP1 Builds
— Build                   Description                           Build              Description
— 10.50.1092        August 2009 CTP2       
— 10.50.1352        November 2009 CTP3
— 10.50.1450        Release Candidate
— 10.50.1600        RTM
— 10.50.1702        RTM CU1
— 10.50.1720        RTM CU2
— 10.50.1734        RTM CU3
— 10.50.1746        RTM CU4
— 10.50.1753        RTM CU5
— 10.50.1765        RTM CU6                    —>    10.50.2500    SP1 RTM
— 10.50.1777        RTM CU7
— 10.50.1797        RTM CU8

 

Also, don’t forget that none of this has anything to do with SQL Server 2008 (which is a separate older, major release according to Microsoft).

Posted in Microsoft, SQL Server 2008 R2 | Tagged | 14 Comments

Teaching Transact-SQL at University of Denver, University College

Starting this week on Wednesday night, I will be teaching a special five-week condensed version of ICT 4462, Transact-SQL Programming on campus at University of Denver, University College. This class is normally a ten week course, but the school usually asks for a condensed version for the Summer quarter. I will cover the same content as the regular ten-week course, just at a faster rate.

We will be using Itzik Ben-Gan’s excellent Microsoft SQL Server 2008 T-SQL Fundamentals book as the textbook. This should be a fun, quick course!

Posted in Teaching | Tagged | Leave a comment

Performance Benefits of Adding RAM to a Laptop

I have a pretty nice Dell Latitude E6420 that I am using as a loaner laptop while I wait for my much nicer Dell Precision M4600 laptop to arrive. The E6420 has a Core i7-2720QM processor, but only had a single 4GB stick of DDR3 RAM installed.

The problem with this is two-fold. First, having only one stick of RAM installed can slow down memory access, since the memory controller in the CPU cannot take advantage of all of the memory channels that it has available. Second, 4GB is really on the low side for a new laptop computer. In mid-2011, I consider 8GB to be a more useful minimum (unless you are really not doing anything to stress your system). Fortunately, DDR3 RAM is quite affordable, so a 4GB stick can be had for $40-50. A quick trip to Micro Center solved that problem!

Looking at the Geekbench scores for the system with 4GB and 8GB of RAM, you see a nearly 50% increase in the Memory and Stream scores with two sticks of RAM vs. one stick of RAM. The overall Geekbench score improved by nearly 13%, which is not bad for less than $50.

More important is the improved ability to run multiple applications at once with 8GB instead of 4GB of RAM. This is much more noticeable in day to day operation.

image

Geekbench Score of Dell Latitude E6420 with 4GB of RAM

 

image

Geekbench Score of Dell Latitude E6420 with 8GB of RAM

Posted in Computer Hardware, Processors | Tagged , | 4 Comments

July 2011 Version of SQL Server 2008 Diagnostic Queries

Since July is only one day away, its time for the updated version of my SQL Server 2008 Diagnostic Information Queries. There is also an updated version of the results spreadsheet, with an additional tab for the latest new query in the set.

This new query lets you see whether SQL Server has any Memory Grants Pending according to the SQL Server: Memory Manager performance monitor counter. If you see any value above zero for any sustained time at all, that is a very strong indicator of internal memory pressure.

These queries will all work on SQL Server 2008 and SQL Server 2008 R2. Most of the queries will also work on SQL Server 2005, but there is a separate set of queries that I have for SQL Server 2005, which I am not updating very often, since SQL Server 2005 went out of Mainstream Support back in April of 2011.

One surprise I have in store quite soon is that I will have two new versions of this script that are split into instance level queries, and database level queries, with even more detail about each area. I am adding quite a few new queries for each new script, so they should be even more useful (at least I hope so)!

Posted in Computer Hardware, SQL Server 2008, SQL Server 2008 R2 | Tagged | 11 Comments

CPU-Z 1.58 Released

Version 1.58 of the very useful utility CPU-Z was released on June 24, 2011. Originally (and still) popular for bragging rights in the PC hardware enthusiast community for proving how much you had been able to overclock a particular processor, it is actually very relevant in a professional I.T. context, primarily for hardware identification.

Running this utility will give you a lot of very detailed information about your processor(s), your motherboard, and your memory. It can also be used to confirm the actual current clock speed of your processor cores, which can be a big issue with some processors (particularly the Intel Xeon 55xx, 56xx,  and 75xx, along with the AMD Opteron 61xx families). when either software or hardware based power management is in effect.

Interestingly, the desktop and mobile Intel Sandy Bridge processors that I have checked so far do not show the huge difference in performance like those earlier processors when they are running with power management enabled.

In case you have never seen or used this utility before, you can see a screenshot of the CPU tab in Figure 1, below.

image

Figure 1: CPU-Z CPU Tab

This utility is free, and in my opinion, completely safe to run on a database server.

Posted in Computer Hardware, Processors | Tagged | 1 Comment

A Tale of Two Sandy Bridge Laptops

I recently bought a new teaching and presentation laptop, which is a Toshiba Portege R835-P55X, which I found at the Microsoft Store in Park Meadows Mall in Lone Tree, CO. This machine has a dual-core 2.3GHz Core i5-2410M, that also has Hyper-threading and Turbo Boost 2.0. It uses the integrated Intel HD graphics, which work perfectly well for business use.

This is a replacement for the very successful Portege R705 line from 2010, which was my previous teaching machine. The new machine looks exactly like my older R705, and it still weighs only 3.2 pounds. The main difference is that it has about 50% more CPU horsepower and significantly better battery life, along with a USB 3.0 port and 6Gbps SATA III support. On the downside, the screen resolution is only 1366 x 768.

The new machine was only $799, but I have about $350 invested in improving it a bit. First, I replaced the two 2GB sticks of RAM with two 4GB sticks of RAM. Second, I put a 6Gbps 128GB Crucial RealSSD C300 solid state drive in it to replace the stock 5400rpm 640GB hard drive. Finally, I put a 32GB PNY Category 10 Micro SHDC card in to have a little more storage space.

After these improvements, here are some metrics and benchmarks on this machine:

image

Windows Experience Index on Toshiba Portege R835-P55X

image

CPU-Z on Toshiba Portege R835-P55X

image

Geekbench on Toshiba Portege R835-P55X

The second machine is a loaner machine that I am using at my new job. It is a Dell Latitude E6420, with a quad-core 2.2GHz Core i7-2720QM, that also has Hyper-threading and Turbo Boost 2.0. This machine only has a single 4GB stick of RAM, and it has 7200rpm 500GB hard drive. It uses the integrated Intel HD graphics, which work perfectly well for business use.

image

Windows Experience Index on Dell Latitude E6420

image

CPU-Z on Dell Latitude E6420

image

Geekbench on Dell Latitude E6420

Having only one stick of RAM really hurts the Latitude E6420 in its Memory scores in Geekbench. In day-to-day usage, having a conventional hard drive (even a 7200 rpm model) makes a huge difference (in a bad way) compared to the Crucial SSD in the little Toshiba Portege. Really, once you have an SSD in a laptop, a conventional hard drive feels painfully slow.

Posted in Computer Hardware, Processors | Tagged | 2 Comments

Presenting Two Sessions at PASS Summit 2011

I found out earlier this week that I was fortunate enough to have my half day session on “Scaling SQL Server” accepted for the PASS Summit 2011, which was quite a pleasant surprise! I had previously been informed that I also had a 90 minute Spotlight Session accepted, which is “Hardware 301: Diving Deeper into Database Hardware”.

Now, I get to actually put together the material and demos for five hours of content on these two subjects (which are at least somewhat related), which will be both a lot of fun and a lot of work! Developing the content for a presentation for a major conference like PASS is a lot of work, but it is something I really enjoy doing. As you do your research, and distill your knowledge and experience into the presentation and demos it tends to be a great learning experience (just like all teaching usually is) that I find to be really valuable.

If the content you develop is good, and if you present it reasonably well, it is very gratifying to find out later that your audience learned something interesting and useful that they were able to take advantage of immediately in their career. That thought makes all the preparation work very worthwhile to me!

Here are the abstracts for the two sessions:

Hardware 301: Diving Deeper into Database Hardware
Enterprise Database Administration and Deployment
Glenn Berry (NewsGator Technologies)

Making the right hardware selection decisions is extremely important for database scalability. Having properly sized and configured hardware can both increase application performance and reduce capital expenses dramatically. Unfortunately, there are so many different choices and options available when it comes to selecting hardware and storage subsystems, it is very easy to make bad choices based on outmoded conventional wisdom. This session will give you a framework for how to pick the right hardware and storage subsystem for your workload type. You will learn how to evaluate and compare key hardware components, such as processors, chipsets, and memory. You will also learn how to evaluate and compare different types of storage subsystems for different database workload types. This session will give you the knowledge you need to make sure you get the best performance and scalability possible from your hardware budget!

Scaling SQL Server
Enterprise Database Administration and Deployment
Glenn Berry (NewsGator Technologies)

How can you scale SQL Server? Scaling up is relatively easy (but can be expensive), while scaling out requires significant engineering time and effort. If you suggest hardware upgrades you may be accused of simply “throwing hardware at the problem”, and if you try to scale out, you may be thwarted by a lack of development resources or 3rd party software restrictions. As your database server nears its load capacity, what can you do? This session gives you concrete, practical advice on how to deal with this situation. Starting with your present workload, configuration and hardware, we will explore how to find and alleviate bottlenecks, whether they are workload related, configuration related, or hardware related. Next, we will cover how you can decide whether you should scale up or scale out your data tier. Once that decision is made, you will learn how to scale up properly, with nearly zero down-time. If you decide to scale out, you will learn about practical, production-ready techniques such as vertical partitioning, horizontal partitioning, and data dependent routing. We will also cover how to use middle-tier caching and other application techniques to increase your overall scalability.

Honestly, the PASS Summit is really the premier SQL Server event, in my opinion. It is a large, well-run event that is very tightly focused on SQL Server. You will be surrounded by a couple thousand people that have to face many of the same issues that you must deal with on a daily basis. People at PASS are very friendly and enthusiastic, which makes the event a lot more fun than some other events. It is definitely a worthwhile investment!

PASS_2011_SpeakingButton_Grey 250x250

Posted in Computer Hardware, PASS, SQL Server 2008, SQL Server 2008 R2, SQL Server Denali, Teaching | Tagged | 1 Comment