Intel P965: Mid-Range Performance Sector Roundup
by Gary Key on October 20, 2006 9:00 PM EST- Posted in
- Motherboards
Disk Controller Performance
The AnandTech iPeak test is designed to measure "pure" hard disk controller performance, and in this case, we keep the hard drive as consistent as possible while varying the hard drive controller. The idea is to measure the performance of each hard drive controller with the same hard drive.
We played back our raw files that are recorded I/O operations when running a real world benchmark - the entire Winstone 2004 suite. Intel's iPeak utility was then used to play back the trace file of all I/O operations that took place during a single run of Business Winstone 2004 and MCC Winstone 2004. To try to isolate performance differences to the controllers that we are testing we use the Seagate 7200.10 Barracuda 320GB 7200 RPM drive in both SATA and IDE offerings for our tests. The drive is formatted before each test run and a composite average of three tests on each controller interface is tabulated in order to ensure consistency in the benchmark.
iPeak gives a mean service time in milliseconds; in other words, the average time that each drive took to fulfill each I/O operation. In order to make the data more understandable, we report the scores as an average number of I/O operations per second so that higher scores translate into better performance. This number is meaningless as far as hard disk performance is concerned, as it is just the number of I/O operations completed in a second. However, the scores are useful for comparing "pure" performance of the storage controllers in this case.
The performance patterns hold steady across both Multimedia Content I/O and Business I/O, with the NVIDIA nForce 500 based SATA controllers providing a slight improvement in I/O operations over the Intel, JMicron, and Silicon Image SATA controllers. This has been a historical trend between the two manufacturers with the NVIDIA controllers generally having a higher throughput than the Intel chipsets. We look forward to seeing how well the ATI SB600 will perform with an Intel processor in the upcoming DFI ICFX-3200 based on the RD600 chipset.
Of note in our SATA tests is the excellent performance generated by the JMicron controller logic that scored higher than the Intel ICH7R or ICH8 based SATA controllers in the Business Winstone test. However, in the read and write heavy Content Creation tests we see the Intel SATA controllers placing ahead of the JMicron offerings. After reviewing the test script results we noticed the write operations of the JMicron JMB363 chipset was about 3% slower than the Intel ICH chipsets while being around 2% faster in read operations.
The VIA VT6410 controller offered the best performance in the IDE tests with the JMicron JMB363 finishing last. During initial testing with the JMicron chipset we noticed several boards having issues with the optical drives being stuck in PIO mode resulting in terrible performance. However, this issue was solved with BIOS updates from the motherboard suppliers. We do want to note that unless you utilize the JMicron JMB363 for RAID operation then do not load the JMicron driver or allow Windows Update to install the newly released driver. Installing this driver will usually result in sporadic operation of the IDE port or failure to recognize your optical drive.
The AnandTech iPeak test is designed to measure "pure" hard disk controller performance, and in this case, we keep the hard drive as consistent as possible while varying the hard drive controller. The idea is to measure the performance of each hard drive controller with the same hard drive.
We played back our raw files that are recorded I/O operations when running a real world benchmark - the entire Winstone 2004 suite. Intel's iPeak utility was then used to play back the trace file of all I/O operations that took place during a single run of Business Winstone 2004 and MCC Winstone 2004. To try to isolate performance differences to the controllers that we are testing we use the Seagate 7200.10 Barracuda 320GB 7200 RPM drive in both SATA and IDE offerings for our tests. The drive is formatted before each test run and a composite average of three tests on each controller interface is tabulated in order to ensure consistency in the benchmark.
iPeak gives a mean service time in milliseconds; in other words, the average time that each drive took to fulfill each I/O operation. In order to make the data more understandable, we report the scores as an average number of I/O operations per second so that higher scores translate into better performance. This number is meaningless as far as hard disk performance is concerned, as it is just the number of I/O operations completed in a second. However, the scores are useful for comparing "pure" performance of the storage controllers in this case.
The performance patterns hold steady across both Multimedia Content I/O and Business I/O, with the NVIDIA nForce 500 based SATA controllers providing a slight improvement in I/O operations over the Intel, JMicron, and Silicon Image SATA controllers. This has been a historical trend between the two manufacturers with the NVIDIA controllers generally having a higher throughput than the Intel chipsets. We look forward to seeing how well the ATI SB600 will perform with an Intel processor in the upcoming DFI ICFX-3200 based on the RD600 chipset.
Of note in our SATA tests is the excellent performance generated by the JMicron controller logic that scored higher than the Intel ICH7R or ICH8 based SATA controllers in the Business Winstone test. However, in the read and write heavy Content Creation tests we see the Intel SATA controllers placing ahead of the JMicron offerings. After reviewing the test script results we noticed the write operations of the JMicron JMB363 chipset was about 3% slower than the Intel ICH chipsets while being around 2% faster in read operations.
The VIA VT6410 controller offered the best performance in the IDE tests with the JMicron JMB363 finishing last. During initial testing with the JMicron chipset we noticed several boards having issues with the optical drives being stuck in PIO mode resulting in terrible performance. However, this issue was solved with BIOS updates from the motherboard suppliers. We do want to note that unless you utilize the JMicron JMB363 for RAID operation then do not load the JMicron driver or allow Windows Update to install the newly released driver. Installing this driver will usually result in sporadic operation of the IDE port or failure to recognize your optical drive.
62 Comments
View All Comments
JarredWalton - Saturday, October 21, 2006 - link
Oh, trust me, Gary tested with a LOT of RAM types and manufacturers. However, for the *benchmarks* he settled on one specific set of DIMMs. I think he's trying to put together some information on how the various boards worked with other RAM (see above comment from Gary). Cheers!stmok - Saturday, October 21, 2006 - link
LOL...I think its more like: "What the hell were the Abit engineers thinking?!"Based on your experiences, do you know if the Analog Devices AD1988A HD Audio Codec works in Linux? I wouldn't mind going for the ASUS P5B-E at the end of the year.
And finally, is the rev 1.02G mobo available now? Or in a few months time?
Gary Key - Saturday, October 21, 2006 - link
The 1.09 ADI AD1988A drivers worked fine in SUSE 10.1. Realtek has better support at this time but ADI seems to be catching up. Believe it or not, but we booted every board with SUSE 10.1 just to make sure they went to the desktop. We might even do a once in a while look at Linux down the road.
xsilver - Saturday, October 21, 2006 - link
what is the range of overclocking possible with pc6400 ram?without dividers? with dividers?
Gary Key - Saturday, October 21, 2006 - link
Good question, depends on the PC2-6400 RAM. We are working on something right now to answer your question with a few different modules.xsilver - Sunday, October 22, 2006 - link
probably something standard like corsair VS or something priced very similar if there is better performance elsewherexsilver - Sunday, October 22, 2006 - link
also I would assume that 1gig vs. 2gig makes no difference but if it does, it would be good to know.Madellga - Saturday, October 21, 2006 - link
Excellent review Gary. I also look forward for the round 2 and also for a 975/ATI/Nvidia reviews.Please try to mention the Vmch used for the overclocking results, as this says a lot about the motherboads also.
Gary Key - Saturday, October 21, 2006 - link
I am setting up a overclocking settings table and will try to show some additional results with different memory types that we used in testing. I have no idea when this will be finished. ;-)Ryan Norton - Saturday, October 21, 2006 - link
This article is terrific-- now I just need to read the high-end part and I can finally put together a Core 2 Duo system.Will the DS4 be included in the high-end guide? I don't want a DQ-6 because the copper backplate under the processor socket would prevent installation of a Scythe Ninja heatsink.
Are the AHCI issues going to stay the way they are, with needing to load the drivers from a floppy disk during Windows install, or are there any P965 boards that don't require "an engineering degree and a day off" as you guys said?