Christian Ney Posted July 8, 2012 Author Posted July 8, 2012 CPU-Z 1.60.x is/was also bugged btw. Last version to work perfect is CPUZ 1.59 but you can't validate with it anymore. But feel free to use this one for screenshots. Know bugged Sockets at the moment for 1.60.x and 1.61.x: LGA775 AM2/2+/3/3+ Quote
just_nuke_em Posted July 9, 2012 Posted July 9, 2012 So I guess these is no point in doing am3+ high freq runs until this is figured out? Quote
I.M.O.G. Posted July 9, 2012 Posted July 9, 2012 GENiEBEN said: Wow, put out 3 bugged versions and block those that still work. Mostly unrelated, but I always had a vague curiosity of why CPUID.com distributed the application, but validation was handled at canardpc.com. Never was curious enough to ask or look into it, but I always thought that was weird. Quote
ScunnyUK Posted July 9, 2012 Posted July 9, 2012 just_nuke_em said: So I guess these is no point in doing am3+ high freq runs until this is figured out? Im putting off ordering LN2 for the time being until this is figured out. You can get a legitimate non bugged screenshot / validation file if you open cpu-z last after overclocking on the fly, only problem is it will still come under scrutiny because no results can really be trusted at the moment. Quote
Christian Ney Posted July 9, 2012 Author Posted July 9, 2012 Okay Franck D. (CPU-Z) did testing the whole day on my system via remote controll from his place. Bugs are fixed (at least the know ones) and seems the precision is close to cpuz 1.59 but not as good (impossible to be as good or better). Why ? Blame microsoft (long story) So I asked him to release it asap and not name it cpuz 1.61. Will double check everything tomorrow but i tested his last version and seems working fine. If you encounter any clocks reading problems or any cpuz related problems with current version, please report here so we can work on. Quote
knopflerbruce Posted July 9, 2012 Posted July 9, 2012 How can it be M$ fault that 1.59 works and this new version is not as accurate? Solution is still simple. Allow both. Quote
Christian Ney Posted July 9, 2012 Author Posted July 9, 2012 (edited) because microsoft use the same api as cpuz in windows 8 therefore they get in conflic that's why cpuz 1.59 and lower were completely bugged in windows 8. Franck D. had to move to another api that seems shit, not very precise and he need to calibrate it. for AMD he did the calibration on Llano that's why Llano is not affected by bug but Llano calibration doesn't work on AM3. This explain the such difference between cpuz 1.6x and cpuz 1.59 and before. Btw all other clocks reading software will have to move to another api as well because microsoft used it and doesn't want to share related stuff. So blame Microsoft Edited July 9, 2012 by Christian Ney Quote
Crew Turrican Posted July 9, 2012 Crew Posted July 9, 2012 interesting. thx for the info chris. Quote
GENiEBEN Posted July 9, 2012 Posted July 9, 2012 (edited) Wait a second, what API are we talking about here? It's safe to assume no WMI is used, so all he has to do is retrieve a specific range from system memory (c/asm) and make use of it. For those interested: http://wiki.osdev.org/Memory_Map_(x86) Christian Ney said: Me no have technical details Nvm https://docs.google.com/open?id=0B7LwI1tJ08joWkd4Ti1seUsySEE Edited July 9, 2012 by GENiEBEN link fixed Quote
just_nuke_em Posted July 11, 2012 Posted July 11, 2012 (edited) nvm, guess he hasn't released it yet Edited July 11, 2012 by just_nuke_em Quote
Christian Ney Posted July 11, 2012 Author Posted July 11, 2012 We have recalibrated CPU-Z for Bulldozer, now we are recalibrating it for AMD K10 Quote
Christian Ney Posted July 12, 2012 Author Posted July 12, 2012 AMD K10 calibrated, now we are doing LGA775 calibration Please let me know if any of you had weird behavior on LGA775 with CPUZ Quote
TaPaKaH Posted July 12, 2012 Posted July 12, 2012 all LGA775 "weird behavior" I've seen so far came from unlocked CPUs Quote
Christian Ney Posted July 12, 2012 Author Posted July 12, 2012 LGA775 calibrated, @ Sam, hum I have to check if I have a LGA775 Unlocked CPU. Quote
Christian Ney Posted July 12, 2012 Author Posted July 12, 2012 (edited) I got the chinese K yeah, remember how to change cpu multi on the fly on LGA775 ? Edited July 12, 2012 by Christian Ney Quote
Eeky NoX Posted July 12, 2012 Posted July 12, 2012 oO Cus it's possible ? euh!? Dunno sorry Long time I haven't booted my lovely E6500K ^^ Quote
Crew Turrican Posted July 12, 2012 Crew Posted July 12, 2012 Eeky NoX said: oO Cus it's possible ? euh!? Dunno sorry Long time I haven't booted my lovely E6500K ^^ no problem, send it to me, i'll take care of it then. Quote
Eeky NoX Posted July 12, 2012 Posted July 12, 2012 Listen to me: NEVEEEEEEEEEEEER!!! Lol I really like that chip but just for fun not for his crappy efficiency of course Quote
TaPaKaH Posted July 12, 2012 Posted July 12, 2012 Christian Ney said: I got the chinese K yeah, remember how to change cpu multi on the fly on LGA775 ? crystal CPUID could do it on the first QX chips, no idea about the E6500K Quote
Christian Ney Posted July 12, 2012 Author Posted July 12, 2012 Thanks, will try, got another soft in mind that could do the job Quote
Christian Ney Posted July 12, 2012 Author Posted July 12, 2012 Managed to reproduce Knut's bug with unlocked CPU, working on the issue right now Quote
Christian Ney Posted July 12, 2012 Author Posted July 12, 2012 Bug on LGA775 fixed now, new build in progress, should be online as soon as possible and valid.canardpc.com will drop 1.60.x and 1.61.x support. New build : 1.61.2 (bug free, at least the know ones) 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.