• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

High CPU utilization with NF7-S, XP Pro SP1, and nForce 2.03 driver pkg. <definite XP SP1 issue> **fix inside**

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
I still don't see the connection and think it is in error in HDTach, not the driver. These tests were only run using one program (HDTach) which is known to be inconsistent and often inaccurate with its CPU tests. I also don't see any benchmarks claiming the driver actually hurts performance (the link you provided makes no mention of a performance increase after the patch), it only seems to post oddly high CPU utilization in one program. MS has posted a patch for the SCSI/ATA controller performance problems in XP and Win2k SP3 finally, but it isn't the one you posted:

Slow Disk Performance When Write Caching Is Enabled
 
Originally posted by: SeekingTao
I wonder if this patch fixes things if you are using the nForce SW IDE drivers? I opted not to install them based on the advice of those who walked before me with the nForce2.

Is there anyone who installed the nForce SW IDE drives and this patch John pointed out that cares to report on this?

I think most people have migrated away from the SW drivers b/c of reported problems with Optical drives as well as the super high CPU utilization, but looks like this patch would drop util. from 60% to 15%! :Q I don't think I even had the option to install the SW drivers in the latest Asus 1.16 UDA (matching nV 2.03).

Chiz
 
Originally posted by: Pariah
I still don't see the connection and think it is in error in HDTach, not the driver. These tests were only run using one program (HDTach) which is known to be inconsistent and often inaccurate with it CPU tests. I also don't see any benchmarks claiming the driver actually hurts performance (the link you provided makes no mention of a performance increase after the patch), it only seems to post oddly high CPU utilization in one program. MS has posted a patch for the SCSI/ATA controller performance problems in XP and Win2k SP3 finally, but it isn't the one you posted:

Slow Disk Performance When Write Caching Is Enabled

I ran a second program and there was a performance increase in that one as well. Look at my post ^^^.

 
Originally posted by: SeekingTao
Originally posted by: Pariah
I still don't see the connection and think it is in error in HDTach, not the driver. These tests were only run using one program (HDTach) which is known to be inconsistent and often inaccurate with it CPU tests. I also don't see any benchmarks claiming the driver actually hurts performance (the link you provided makes no mention of a performance increase after the patch), it only seems to post oddly high CPU utilization in one program. MS has posted a patch for the SCSI/ATA controller performance problems in XP and Win2k SP3 finally, but it isn't the one you posted:

Slow Disk Performance When Write Caching Is Enabled

I ran a second program and there was a performance increase in that one as well. Look at my post ^^^.

I just went and downloaded HD_Speed and tried it. The graph gives you no indication of where on the drive it is testing. Assuming it starts at the beginning of the drive (the fastest part) the measurement wasn't accurate for any of the 6 drives in my system. I'll also note, that the 48MB/s you got using HD_Speed is still about 5MB/s too slow for that drive. I would not use that program, it doesn't look accurate.

Hard Tecs 4U

If you look at the link above to the original NForce review that noticed the high CPU (linked off the link John provided from Warp), you will also notice that everything except the CPU numbers are accurate (STR beg/end/average, access time, burst) so whatever is causing the odd CPU measurement is not affecting drive performance.
 
dayum, my system seem fast enough for me that I didn't complain but after running hdtach, i noticed my cpu utilization was a whooping 55%!!! actually now that i think about it, right clicking seem sluggish on my system. but after doing the fix my cpu usage is less than 10% and right clicking is instantaneous, nice post.
 
alright I installed the fix, and my utilization went from 50% to about 8%, but my sandra score went down from around 31k to 28.5k, and with hd tach, my seek time went up from 13.3 to 14ms
 
I installed this patch immediately when building a NF2 boxen as I followed the 6P's. well I forgot I had to do a repair install after nuking the system32/config file among others while getting too aggressive overclocking. I started having a slooooow boot issue and Chiz suggested disabling the SATA option in bios and that helped a little but it was still there, I ran across this thread and the light when on so I installed the patch again and slooooow boot gone 🙂 Bump!
 
Whoa! Just did that fix on my KT333 system and the CPU utilization went down from around 50 on both of my drives to 11 and 8! AWESOME!
 
Well, I did this on my KT400-based system and I'm not sure if it fixed my issue of skipping occaisonally on DVDs and music (mp3s or CDs) so i'll leave it on for now. The CPU utilization is at about 0-7% and i'm running AIM/Internet Explorer (obviously), and Winamp 2.81. It reduced my score in Sandra from about 35Megabytes/sec to 25megabytes/sec, which really surprised me. Sandra isn't a good way to gauge Hard Drive performance, in my opinion, so I'll just sit and see if it really is slower. I'll trade such a performance decrease, however, if my skipping issue is resolved, cuz it sure was a pain when it occurred.
 
Back
Top