Separate names with a comma.
Discussion in 'CPUs and Overclocking' started by polyzp, May 2, 2012.
+1. I can do Phenom2 x4 at 4ghz. When you limit to 4 threads it shouldn't be any worse than Phenom2.
Awesome man! I have a few requests for you then!
Fritz Chess 4.3 - 1 core performance
Cinebench 10 - 1 core performance
Cinebench 11.5 - 1 core performance
TechARP x264 HD - first pass
would love to see how a single 4.9 ghz FX thread goes up against a 4.0 Ghz phenom II thread!
Thanks again man! Ill use your results in my next post! ^^
And there is the daily bump to make sure it never falls off the first page, just like clockwork!
Paid by AMD, for sure. But hey, this way we will never forget what a lame duck bulldozer is.
Ye this guys been a joke since day 1.
I wonder what cost him more by now. His faildozer or the powerbill it creates
Well, at least he can always count on his heart being warmed by the heat of your Ivy
polyzp: maybe u can run 4m/4t at max oc and at 4ghz to compare vs the phenom x4?
Yes Im open to this idea. Im mostly interested in comparing single thread performance for this comparison.
It would take an act of God to make it happen in the first place.
the single core ipc for BD v Ivy in cinebench is at a 40% deficit running at about same speed. That's crazy poor.
But Ivy is at a 50% deficit on number of cores, so BD is actually faster when 8 threads are used. See how useless "ipc per core" is as a number?
Which part? Do you dispute that bulldozer is 8 cores, or do you dispute that 4 is 50% of 8? Or do you disagree that bulldozer is 40% slower for a single core IPC?
That is one sexy board
Scaling for bulldozer isnt double the 2600k, but neither is IPC of ivy over FX (although in some cases it might be, generally it is not). When all threads are used it just about makes up for lost IPC per thread. FX is seen to fail most when all 8 threads arent taken advantage of.
Incorrect. I guess you suddenly forgot that Hyper-Threading is a 20% improvement in MT.
And Cinebench is mostly floating-point code, and Bulldozer has four FPUs.
cb10-- 4275 1 core
4 core 15704
Results for x264.exe v0.58.747
encoded 1442 frames, 76.16 fps, 3904.67 kb/s
encoded 1442 frames, 75.59 fps, 3904.67 kb/s
encoded 1442 frames, 76.79 fps, 3904.67 kb/s
encoded 1442 frames, 77.39 fps, 3904.67 kb/s
encoded 1442 frames, 20.90 fps, 3952.97 kb/s
encoded 1442 frames, 20.71 fps, 3952.97 kb/s
encoded 1442 frames, 20.76 fps, 3952.97 kb/s
encoded 1442 frames, 20.94 fps, 3952.97 kb/s
Results for x264.exe v0.59.819M
encoded 1442 frames, 77.99 fps, 3889.34 kb/s
encoded 1442 frames, 78.29 fps, 3889.34 kb/s
encoded 1442 frames, 78.15 fps, 3889.34 kb/s
encoded 1442 frames, 78.14 fps, 3889.34 kb/s
encoded 1442 frames, 22.41 fps, 3962.75 kb/s
encoded 1442 frames, 22.11 fps, 3962.75 kb/s
encoded 1442 frames, 22.14 fps, 3962.75 kb/s
encoded 1442 frames, 22.06 fps, 3962.74 kb/s
I'm running 15.5x260 = 4,030mhz
Power is only of no concern when money is not an object, and I would not get a bulldozer chip if money were not an object, unless it was to do a very specific workload that bulldozer excels in.
Unfortunately money typically is an object for most people, and even if BD is faster in a particular workload, it might be more cost effective over time to pony up for a dual Intel system.
I dislike the thought of Intel being able to sell their server chips at a much higher price despite having smaller dies, because companies will still save money over the product's lifetime compared with AMD chips.
Thanks man, is that tech arp x264 HD? the scores seem off
heres a link
and whats your single core performance in cb11.5?
apparently not redownloaded and ran--
encoded 1442 frames, 103.23 fps, 3912.32 kb/s
encoded 1442 frames, 105.09 fps, 3912.32 kb/s
encoded 1442 frames, 105.57 fps, 3912.32 kb/s
encoded 1442 frames, 105.02 fps, 3912.32 kb/s
encoded 1442 frames, 25.10 fps, 3961.76 kb/s
encoded 1442 frames, 25.14 fps, 3961.07 kb/s
encoded 1442 frames, 25.22 fps, 3961.36 kb/s
encoded 1442 frames, 25.16 fps, 3961.01 kb/s