OBR Posted November 19, 2011 Posted November 19, 2011 (edited) Big error in ALL reviews (mine too) ... AMD FX-8150 King of WinRAR! Real results! after Power management OFF! Change it in there: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Power\PowerSettings\54533251-82be-4824-96c1-47b60b740d00\0cc5b647-c1df-4637-891a-dec35c318583 (ValueMax from 64 to 0) and enjoy WinRAR ride on Intel HT CPUs! Edited November 19, 2011 by OBR Quote
OBR Posted November 19, 2011 Author Posted November 19, 2011 this register is disabling Core Parking in Windows 7 ... there is BUG for Hyper-Threading CPUs ... When is enabled (value 64) HT cores are sleeping ... in WinRAR (look at resource monitor NOT performance monitor), when disabled ALL cores are fully working ... and Multi-thread results in WinRAR are great! Quote
Hondacity Posted November 20, 2011 Posted November 20, 2011 let the whole world know.. nice work OBR Quote
Le085 Posted November 20, 2011 Posted November 20, 2011 We would confirm your findings, OBR. We also noticed that the application is affected too, not only the WinRAR benchmark. Compressing an ISO with high compression settings, the CPU load jumped from about 30% to over 50%. http://translate.google.it/translate?hl=it&sl=auto&tl=en&u=http%3A%2F%2Fwww.xtremehardware.it%2Fnews%2Fhardware%2Fobr-sveglia-i-core-dormienti-in-sandy-bridge-e-rivela-le-reali-prestazioni-in-winrar-201111206080%2F Nice job!! Thank you for sharing. Quote
flanker Posted November 21, 2011 Posted November 21, 2011 Hi Zdenek, of course, this is right, but we can not change OS settings in reviews, if is it fault of Windows7,...then that is it. Because we can than testing with affinity at cores etc etc. and all results will be little changed. It is really good for daily using and in reviews as interest ing fact. Quote
OBR Posted November 21, 2011 Author Posted November 21, 2011 of course, this is BS man. If you want to review anything, you HAVE TO unveil full potential and show it to readers. These first results were wrong, now i have this feature DISABLED forever in CPU reviews HDD ... Quote
Massman Posted November 21, 2011 Posted November 21, 2011 Heh. Funny how some people forced reviewers to also test Windows8 because that scheduler would unleash "full potential" of Bulldozer. But now that a simple reg key can unleash SBE power, it's not allowed to use it . Quote
Hondacity Posted November 21, 2011 Posted November 21, 2011 Hi Zdenek, of course, this is right, but we can not change OS settings in reviews, if is it fault of Windows7,...then that is it. Because we can than testing with affinity at cores etc etc. and all results will be little changed. It is really good for daily using and in reviews as interest ing fact. you're a certified amd biased fan Quote
Crew Leeghoofd Posted November 21, 2011 Crew Posted November 21, 2011 I fail to see the reason in your reasoning Flanker... Quote
OBR Posted November 21, 2011 Author Posted November 21, 2011 Heh. Funny how some people forced reviewers to also test Windows8 because that scheduler would unleash "full potential" of Bulldozer. But now that a simple reg key can unleash SBE power, it's not allowed to use it . Bulldozer + Windows8 - fail like Bulldozer + Windows7 ive measured it ... http://pctuning.tyden.cz/hardware/procesory-pameti/22476-bulldozer-v-prostredi-windows-8-srovnani-vykonu?start=13 Quote
rebelion Posted November 21, 2011 Posted November 21, 2011 I´m agree with flanker because not everybody know how to do that tweak. Lot of people who read the reviews don´t want to touch their system, so if you don´t tweak this key everybody will see the real performance with a retail OS. Did you tweak your system in the same way for reviews and for overclocking??? that´s the point i think. But you did a really good job OBR! See you guys! Quote
flanker Posted November 21, 2011 Posted November 21, 2011 Massman: It is simillary as this idea: IMC of SB 1155 is 1333 MHz, no more (if u want believe me, download datasheet). FX has 1866 MHz default IMC. So than why people tetsing both platforms with 1600 MHz usually? Or....Some games and again FX, we know, FX scaled better with core 0/2/4/...Give priority to these cores and u will se "big" diferences. Its fair? Its cheat? No, it is the same as Windows7 and parking threads. Windows 7 is not perfect, but we are not Microsoft, so average user John dont will have tweaking windows7. Most people put CPU in to MB, installing OS and thats all. We are diferent people here in forums.... Or looks at some boards and Turbo ratio. SOme boards for SB-dt has classic Intel turbo at default setings-, but some boards performed more better, because here is modified turbo (some ASUS and Gigabyte boards). If is somethink default, I testing with this default, no changing it for "review". For interesting value I can show people some tweaks etc, of course, this is right. This thought is not about my AMD or Intel fans.... Quote
I.M.O.G. Posted November 21, 2011 Posted November 21, 2011 The most important part of a review: Is it verifiable? In science, that is the key to testing and proving things - can the results be reproduced/verified? In reviews, if you say how the test was performed and what the result was - anyone using the same platform should have all the information necessary to reproduce similar results, verifying the accuracy of your results. If you are tweaking the default OS install in order to modify the results you achieve in a review, this should be clearly stated in the review. If it is, you are good to go. Many people are not aware of tweaks like this which optimize for multi-threaded extraction/compression performance under HT. If Windows hasn't released a fix to actually resolve the reported bug, I'd suggest that the performance results given from manually tweaking the registry could give most people the wrong impression about the performance they are likely to see from the processor in their system at home... Unless they are willing to edit the registry as you describe in your review. Do you know of any power usage impact? The change disables core parking, which is described here as a power saving feature. Quote
Hondacity Posted November 22, 2011 Posted November 22, 2011 The most important part of a review: Is it verifiable? In science, that is the key to testing and proving things - can the results be reproduced/verified? In reviews, if you say how the test was performed and what the result was - anyone using the same platform should have all the information necessary to reproduce similar results, verifying the accuracy of your results. If you are tweaking the default OS install in order to modify the results you achieve in a review, this should be clearly stated in the review. If it is, you are good to go. Many people are not aware of tweaks like this which optimize for multi-threaded extraction/compression performance under HT. If Windows hasn't released a fix to actually resolve the reported bug, I'd suggest that the performance results given from manually tweaking the registry could give most people the wrong impression about the performance they are likely to see from the processor in their system at home... Unless they are willing to edit the registry as you describe in your review. Do you know of any power usage impact? The change disables core parking, which is described here as a power saving feature. microsoft will prolly release a patch for it... Quote
I.M.O.G. Posted November 22, 2011 Posted November 22, 2011 microsoft will prolly release a patch for it... I don't think there is any basis upon which to suggest or believe that. It looks like the core parking issue has been well documented across the internet since October 2009, for example: http://forum.cakewalk.com/tm.aspx?high=&m=1852473 Like you, I would like to think Microsoft would patch something like this - but they haven't and I don't know if that is because they don't consider it a bug, or maybe it saves power, or maybe they don't care. Dunno, but after it being a known issue for over 2 years, I wouldn't expect them to release a fix for it. Quote
flanker Posted November 22, 2011 Posted November 22, 2011 HondaCity: I dont think it....It will be maybe in windows8, Windows7 is "dead" for Microsoft, they focus for win8 last months. Quote
Crew Leeghoofd Posted November 22, 2011 Crew Posted November 22, 2011 (edited) I agree with Flanker , prolly no patch will appear for any of the CPUs. I even doubt win8 will be even it... Edited November 22, 2011 by Leeghoofd Quote
Massman Posted November 22, 2011 Posted November 22, 2011 Isn't the whole point of being an enthusiast that we try to figure out bottlenecks and work around them if necessary? CPC comparisons at overclocked speeds, max clock performance comparisons, performance scaling articles ... they're all meant to have greater insight in how the platform behaves and how it works. Most power users find some issues and try to solve them. For instance, I spend a lot of time on Bulldozer 3DMark11 to find out there's something strange with memory divider and memory performance. OBR saw there is something strange in WinRAR with the Intel CPUs, so he found out what the bug was. That's just what being an enthusiast is all about. For regular people, I'm sure someone will just make a patch and people will use the patch if it gives massive performance gain. In the past, we've all updated our OS with patches to gain performance ... Massman: It is simillary as this idea:IMC of SB 1155 is 1333 MHz Link? Quote
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.