Investigations into Athlon X2 Overclocking
by Jarred Walton on December 21, 2005 12:00 PM EST- Posted in
- CPUs
RAM Latency
In the Venice article, we talked quite a bit about memory latencies and how they affect performance. Since we're using four different types of RAM this time, we will most likely see more variation in RAM latencies. However, even a relatively large difference of 20% will often have less than a 5% impact on real world application performance. Another item to pay attention to is the difference between higher bandwidth vs. lower latencies that the OCZ VX and EL Platinum represent.
CPU-Z includes a latency.exe program that we used to get these values, and we selected the score in the bottom-right of the table. This position represents a 32MB data set with a 512 KB stride size, and the results are reported in CPU cycles. We have two charts again, the first in CPU cycles and the second in nanoseconds. Since cycle time decreases as clock speed increases, we would pay a bit more attention to the results in nanoseconds. However, the elapsed CPU cycles are also important as they represent wasted CPU time. If cache misses are relatively constant - and they are - then the higher the RAM latency is in CPU cycles, the less efficient the processor becomes. Here are the results.
Since all of the various configurations ran successfully with 1T timings, latencies stay relatively close - even the value RAM isn't that far behind the fastest RAM used. We couldn't get results for 4x512MB except at stock clock speeds, but comparing the scores at 10x200 is still possible. You can clearly see how much the 2T command rate impacts performance, which brings up another point: a lot of value RAM is not guaranteed to run at 1T command rates. This same value RAM required 2T command rate with the Venice chip for the 9x300 setting, even though it was still running below its rated DDR400 speed. Your mileage may vary.
Something else that we failed to mention last time was the numerous in-between options for RAM. 2.5-3-3-7/8 RAM can be purchased for around $80, give or take. In the $100 to $130 price range, there are many sets of 2x512MB DIMMs rated for 2-3-3-6-1T or 2-3-2-5-1T timings. The difference between CL2 and CL2.5 isn't huge, but it may warrant spending an extra $20. The true performance/overclocking RAM starts at around $120 (Crucial Ballistix - really an awesome deal if you ask me!) and goes up from there to over $200. That's a dramatic price increase on the high-end for what is likely to be a small performance difference, but there is something to be said for the ease of use that expensive RAM offers. Where we had to do some work finding the optimal performance settings on value RAM at the various overclocked speeds (and we really only scratched the surface), the OCZ EL Platinum could usually be left at a DDR-400 setting in the BIOS with 2.5-3-3-7-1T timings. It wasn't optimal in all cases, but it required very little effort to find a stable overclock.
The OCZ VX is definitely held back by the Infinity motherboard. 3.2V is the minimum required voltage for stable 2-2-2-8-1T timings. We were able to reach as high as DDR-452 speeds while maintaining 2-2-2-8 timings, but DDR-460 required 2-3-2-8-1T and DDR-500 required 2-3-3-8-1T timings. Beyond DDR-500, even raising CL to 2.5 didn't help much. OCZ VX is still very fast RAM, but it really needs the best in terms of motherboards to reach its full potential. In the RAM area, we now have two strikes against our selected motherboard.
In the Venice article, we talked quite a bit about memory latencies and how they affect performance. Since we're using four different types of RAM this time, we will most likely see more variation in RAM latencies. However, even a relatively large difference of 20% will often have less than a 5% impact on real world application performance. Another item to pay attention to is the difference between higher bandwidth vs. lower latencies that the OCZ VX and EL Platinum represent.
CPU-Z includes a latency.exe program that we used to get these values, and we selected the score in the bottom-right of the table. This position represents a 32MB data set with a 512 KB stride size, and the results are reported in CPU cycles. We have two charts again, the first in CPU cycles and the second in nanoseconds. Since cycle time decreases as clock speed increases, we would pay a bit more attention to the results in nanoseconds. However, the elapsed CPU cycles are also important as they represent wasted CPU time. If cache misses are relatively constant - and they are - then the higher the RAM latency is in CPU cycles, the less efficient the processor becomes. Here are the results.
Since all of the various configurations ran successfully with 1T timings, latencies stay relatively close - even the value RAM isn't that far behind the fastest RAM used. We couldn't get results for 4x512MB except at stock clock speeds, but comparing the scores at 10x200 is still possible. You can clearly see how much the 2T command rate impacts performance, which brings up another point: a lot of value RAM is not guaranteed to run at 1T command rates. This same value RAM required 2T command rate with the Venice chip for the 9x300 setting, even though it was still running below its rated DDR400 speed. Your mileage may vary.
Something else that we failed to mention last time was the numerous in-between options for RAM. 2.5-3-3-7/8 RAM can be purchased for around $80, give or take. In the $100 to $130 price range, there are many sets of 2x512MB DIMMs rated for 2-3-3-6-1T or 2-3-2-5-1T timings. The difference between CL2 and CL2.5 isn't huge, but it may warrant spending an extra $20. The true performance/overclocking RAM starts at around $120 (Crucial Ballistix - really an awesome deal if you ask me!) and goes up from there to over $200. That's a dramatic price increase on the high-end for what is likely to be a small performance difference, but there is something to be said for the ease of use that expensive RAM offers. Where we had to do some work finding the optimal performance settings on value RAM at the various overclocked speeds (and we really only scratched the surface), the OCZ EL Platinum could usually be left at a DDR-400 setting in the BIOS with 2.5-3-3-7-1T timings. It wasn't optimal in all cases, but it required very little effort to find a stable overclock.
The OCZ VX is definitely held back by the Infinity motherboard. 3.2V is the minimum required voltage for stable 2-2-2-8-1T timings. We were able to reach as high as DDR-452 speeds while maintaining 2-2-2-8 timings, but DDR-460 required 2-3-2-8-1T and DDR-500 required 2-3-3-8-1T timings. Beyond DDR-500, even raising CL to 2.5 didn't help much. OCZ VX is still very fast RAM, but it really needs the best in terms of motherboards to reach its full potential. In the RAM area, we now have two strikes against our selected motherboard.
Facebook
Twitter
RSS
46 Comments
View All Comments
JarredWalton - Wednesday, December 21, 2005 - link
Ugh - at a comment on is an article that true special attention to the fact that the graphs aren't zeroed. I think in the process of tweaking article to get things to look right, I accidentally deleted that paragraph. I have now http://www.anandtech.com/cpuchipsets/showdoc.aspx?...">added to paragraph back in.If I start all the graphs at zero, everything overlaps and you can't really see what's going on. In some cases, everything is still overlapped a lot (FEAR). I normally hate nonzero graphs, but when the results are all so close together, that's no good either for readability.
BigLan - Wednesday, December 21, 2005 - link
Well, if everything is overlapping that much, it's likely that the results are too close to be really meaningful. The FEAR graph is a pretty good example of <a href="http://www.amazon.com/gp/product/0393310728/>Ho...">http://www.amazon.com/gp/product/0393310728/"... to Lie With Statistics</a> ;)BigLan - Wednesday, December 21, 2005 - link
^ Still need my edit function for comments. :pDammit
JustAnAverageGuy - Wednesday, December 21, 2005 - link
:)JarredWalton - Wednesday, December 21, 2005 - link
^ Still need my edit function for comments. :pThat was supposed to say, "I had a comment in this article that drew special...." That will teach me to trust my speech recognition software.
Hacp - Wednesday, December 21, 2005 - link
I have found that past 2.6, the heat and temps increase dramatically. Nice to know that anandtech got the same results as well.Yianaki - Wednesday, December 21, 2005 - link
SuperPI crashes help! I have a Opteron 165 with 1 gig of value PC4000 kingmax ram at 133 2.5-4-4-8 2t. Board is OC to 1.4v at 9x277 = 2.494. I have run two torture versions of prime 95 (one of the CPU intensive, one of the ram intensive) on each processor, for a total of 4 prime95's. At the same time I run 3dmark 2005. At the same time I run winamp with visualizations on. I leave this on all night 9hrs+ in a loop. No errors at all. No buggyness at all. I game for 3 or 4 hours at a time and no problems.But I just read this article and SuperPI was mentioned and I never used that before and I tried it. It wouldn't work unless I lowered my overclock to 2.00 which is unacceptable to my sorry ass. I KNOW my system is unstable. I just was wondering if it mattered as the computer is completely stable. I am guessing that prime95 just rounds off answers and they don't have to be exact whereas I am guessing that SuperPI's answers are already known to be exact. Actually SuperPI runs fine but if I open up a second copy from a second folder and attach the affinity to both processors SuperPI will have errors as soon as I start it 1second of starting the 2nd process. Any ideas??
Could it possibly be my motherboard or ram as both are 'value' versions not OC specialty items. I have already played around with rendering divx movies and playing doom no problem. Will I probably have some problem down the road or like some small encoding error or dvd writing error that is due to my overclock. I Overclocked my old PII too high and it was spewing out bad math. I did all these chem reports in college and was getting completly off the wall numbers (I never tested my PII oc in prim95). Is this the same or does the error correcting in my programs that I am running in XP make this point moot.
Man I wish I never read about superPI poo :<
Furen - Wednesday, December 21, 2005 - link
Superpi and prime95 work differently. I think superPI is more reliant on memory bandwidth (if you're doing something like 32M superPi) so your ram may be the problem (I, personally, like crucial ram, I've never had any problems with it at all, and its not much more expensive than the generic crap). If your system doesnt crash when you're doing whatever it is that you do on your computer then you're fine, though, but I'd still try to work on the ram to see if you can get it to be superpi stable.Yianaki - Wednesday, December 21, 2005 - link
It just gets wierder and wierder... If I don't set the affinity manually in the task manager it will run through to the end and the CPU's will both be at 100%. But if I set them manually in the task manager before I start 32M the second one I start always crashes? I am guessing that the task manager isn't running the processors at 100% or something, as the windows task manager is automatically putting the loads on the two cpus and for a milisecond one isn't doing anything???? My Memory is up to 95% utilization... This proggie sucks if you ask me.Yianaki - Wednesday, December 21, 2005 - link
Thanks for the feedback. BTW it is Kingmax Super Ram, Dual channel mate. My motherboard is a ASRock 939 dual sata-2. Bought it because it runs AGP and PCIe quickly. I have my ram underclocked normally it can run at 200 but is running at 133 normally, I also lowered it to 100. I lowered all the timings lower than what the spd says. None of these things help... I am really confused. I run the Blend (ram intensive) test on each processor at the same time as the FFT test in prime 95. Memory usage goes up to 1.5 gigs total (I only have a gig), so that is using all the memory + page. But there is no error at all. I am a little dumbfounded but I have been thinking about it and my computer doesn't have any 'random' errors which is fine. Cept for firefox 1.5 and it had the same occasional problems before my upgrade. Oh well hope everything stays stable. Thanks for the feedback.