Domain: tpc.org
Stories and comments across the archive that link to tpc.org.
Stories · 13
-
Why Don't Open Source Databases Use GPUs?
An anonymous reader writes "A recent paper from Georgia Tech (abstract, paper itself) describes a system than can run the complete TPC-H benchmark suite on an NVIDIA Titan card, at a 7x speedup over a commercial database running on a 32-core Amazon EC2 node, and a 68x speedup over a single core Xeon. A previous story described an MIT project that achieved similar speedups. There has been a steady trickle of work on GPU-accelerated database systems for several years, but it doesn't seem like any code has made it into Open Source databases like MonetDB, MySQL, CouchDB, etc. Why not? Many queries that I write are simpler than TPC-H, so what's holding them back?" -
CowboyNeal Reviews Oracle Linux
CowboyNeal writes "Last week, Oracle announced that they were making Oracle Linux available free of charge, and also provided a script that makes switching to Oracle Linux nearly painless for existing CentOS users. What makes Oracle Linux unique, and why would anyone want to use it? Read on to find out, as I delve into what Oracle Linux has to offer."What is Oracle Linux?
On its face, Oracle Linux feels like just another Red Hat Enterprise Linux (RHEL) derivative. It uses anaconda for an installer. It uses yum for handling packages. Configuration is handled just like RHEL, CentOS, or Scientific Linux. To be honest, the reasons why anyone would switch to Oracle Linux aren't immediately apparent after installing. It feels like nearly any other Linux with the Oracle name bolted on. Under the hood, however, are some rather compelling features.
The Unbreakable Enterprise Kernel
I have to start off with saying that I hate the name "Unbreakable Enterprise Kernel." I've seen enough crazy stuff in my time, to know that no software is truly unbreakable. It might be pretty good, but unbreakable is like calling the Titanic unsinkable. Given a poor enough captain, or in this case, an administrator, I don't have any doubts that the kernel could be broken in at least some fashion. However, I suppose that "Unbreakable Enterprise Kernel" sounds a lot better than the "Pretty-dang-tootin'-robust Enterprise Kernel," and with a target like enterprise customers, terms like "Pretty-dang-tootin'" just won't get stuffy execs to authorize using Oracle Linux.
With that off my chest, let's see what the Unbreakable Linux Kernel does have to offer. Oracle has added a number of their own enhancements into a Linux 2.6 kernel. These include networking optimizations, NUMA optimizations, and enhancements for OCFS2, asynchronous I/O, SSD disk access, OLTP, and more. They clearly work pretty well, as back in March, Oracle submitted a TPC-C benchmark for a Sun Fire server that was the fastest x64-based non-clustered system.
Ksplice: Update Your Kernel Without Rebooting
Ksplice was acquired by Oracle roughly a year ago, and as a result is married to Oracle Linux rather nicely. Ksplice is the holy grail for any administrator who is obsessed with uptime. It gives you the ability to update your kernel, with no downtime necessary. This is by far the best reason to use Oracle Linux, but it also comes at a steep price. While support for Ksplice is present in the Unbreakable Enterprise Kernel, it does nothing without the Ksplice Uptrack service enabled.
How does one get Ksplice Uptrack? It's only included with an Oracle premier support contract. While Oracle is quick to note that it costs less than a similar-tier RHEL support contract, it's also still more than most people would want to pay for just reboot-less kernel updates. Sure, there's also actual support included in the contract, but the lack of an ala carte option for just Ksplice Uptrack doesn't make a premier support contract any easier to swallow.
I should note here, that regular package updates via yum, and regular kernel updates via yum, are still totally free. If you don't mind rebooting, Ksplice isn't a must-have. If Oracle wanted to attract more customers, an ala carte option for Ksplice Uptrack would be a step in the right direction. If they wanted to really build some good will with the Linux community, they'd make Ksplice Uptrack free for everyone. I know it may sound weird to mention Oracle and good will together, but I'd never thought I'd see Oracle and "free" mentioned together either. As it is, it feels like Uptrack is being used as the bait for a support contract, when the support contract should really be able to stand on its own.
DTrace: Debugging and Troubleshooting in Real Time
To be fair, the DTrace modules can be plugged into a lot of Linux kernels already out there, but Oracle Linux has done the leg work for their users. Maybe you're not doing the sort of development that requires DTrace, but it's still something handy to have in the toolbox when something breaks. It's also a handy way to profile already running processes at any moment, with little to no impact on performance when tracing a process. Oracle maintains a long list of DTrace resources on their OpenSolaris site.
Should I give this a look?
If you're already perfectly happy with your RHEL or CentOS Linux install, Oracle Linux is a hard sell, even at the price of free. After toying about with the system, I'd say it's at least worth a hard look. As it is, you get the benefits of CentOS or Scientific Linux, with Oracle's own stuff bolted on, and their enhancements, even minus Ksplice, make a compelling argument to use Oracle Linux. If you are setting up a machine to use Oracle's database software, Oracle Linux is the best choice, since it's been designed to support Oracle DB, and is the same Linux that Oracle uses in-house. While Oracle's premier support contract is cheaper than the RHEL alternative, the actual cost of switching from RHEL to Oracle in a given case may not be. While this release is a good first step for Oracle, more options, like free Ksplice Uptrack, or even a Ksplice Uptrack subscription, would make it an easier sell.
If you'd like to give Oracle Linux a try, without having to jump through a lot of hoops, the Oracle Linux Wiki has a list of download sites.
-
New Linux TPC-H Record Set
prostoalex writes: "New TPC-H world record for performance and scalability of database software on Linux platform has been set. The winner - Oracle 10g running on a four-node Lenovo Cluster Server DeepComp 6800, each with four Intel Itanium 2 1.3 GHz processors. Oracle also emphasizes that it's 3.5 times more performance than similar IBM DB2 benchmark. TPC-H benchmarks are available at TPC Web site." -
First Industry-Standard Benchmark On 64-bit Linux
Haider writes "IBM has published two TPC-H benchmarks at the 100GB and 300GB scale factors (meaning the raw data in the database). The results are published at www.tpc.org/tpch/results/tpch_perf_results. The two results were published with exactly the same hardware and show scalability of the solution as more data is put on it. The system was a cluster of 8 2-way nodes running 2GHz AMD Opteron cpus with 6 GB of memory on each node. The OS was SLES 8 for AMD, and the database was 64-bit DB2 for Linux. The details of this solution are described in an article at infoworld.com." -
OSDL Releases TPC Benchmark Tests For Linux
trialsboy writes "A news.com article announces that ODSL has released a set of workloads and test frameworks which conform to the TPC benchmark specifications." (Slashdot parent company VA Software is one of the OSDL sponsors.) -
HP Publishs First Linux TPC-C Benchmarks
The first ever official TPC-C benchmark on a Linux system has been published. This was run on a cluster of 32 HP servers with Intel Xeon CPUs, running Redhat Linux and Oracle RDBMS. The system had over 18 terabytes of storage, and cost over 2 million US dollars. Performance was higher than a similar system running on MS Windows. -
HP Publishs First Linux TPC-C Benchmarks
The first ever official TPC-C benchmark on a Linux system has been published. This was run on a cluster of 32 HP servers with Intel Xeon CPUs, running Redhat Linux and Oracle RDBMS. The system had over 18 terabytes of storage, and cost over 2 million US dollars. Performance was higher than a similar system running on MS Windows. -
HP Publishs First Linux TPC-C Benchmarks
The first ever official TPC-C benchmark on a Linux system has been published. This was run on a cluster of 32 HP servers with Intel Xeon CPUs, running Redhat Linux and Oracle RDBMS. The system had over 18 terabytes of storage, and cost over 2 million US dollars. Performance was higher than a similar system running on MS Windows. -
Linux Grabs World Record For TPC-H Benchmark
An Anonymous Coward writes: "Linux 2.4.3 now holds the world record by performance with IBM's DB2 in TPC-H. TPC-H is a decision support benchmark consisting of a suite of business oriented and ad-hoc queries and concurrent data modifications. This is way cool as the world record was held by SQL Server 2000 on Windows 2000 before." Caveats: this is only in the 100GB (smallest) category, and all but 2 of the other entries are several months old. Even so;) -
Fujitsu SPARC64-GP vs. Sun UltraSPARC II?
AtariDatacenter asks: "I've noticed that Fujitsu is starting to market their Primepower line of Sun compatible servers, based on the SPARC64-GP processor. Their Primepower 2000, with 128 processors running at 563mhz, would seem to be a killer of the Sun E10000, which is based on the UltraSPARC II processor. And their server beats the Sun E10k in the top ten TPC-C performance results for non-clustered systems. What do Slashdot readers have to say about the Primepower line of servers?" -
TPC-C Benchmarks For JDBC?
woggo asks: "I need to benchmark two different JDBC drivers for a research project and would like to use a standard benchmark. I was able to find this implementation of TPC-W, but that is too much of a test of the Web server to be useful for my purposes. Does anyone know of a freely-available Java implementation of TPC-C? It needs to be reasonably conformant and I need to be able to cite the results in a paper without violating a license agreement, which would seem to exclude evaluation versions of products." -
Are TPC Benchmarks A Worthwhile Measure?
KhaosSpawn asks: "Microsoft is back at the top of the TPC-C performance results and has all ten spots in the top ten by price category. But how many of you work for companies that own a ProLiant 8500-700-192P machine? Are there any results for 2 and 4 way machines? Are there any statistics which include Linux installations? Does this TPC figure mean anything in terms of real world deployment or is this just another number for the Microsoft Marketing Machine?" -
Are TPC Benchmarks A Worthwhile Measure?
KhaosSpawn asks: "Microsoft is back at the top of the TPC-C performance results and has all ten spots in the top ten by price category. But how many of you work for companies that own a ProLiant 8500-700-192P machine? Are there any results for 2 and 4 way machines? Are there any statistics which include Linux installations? Does this TPC figure mean anything in terms of real world deployment or is this just another number for the Microsoft Marketing Machine?"