ludek Posted September 29, 2015 Posted September 29, 2015 Wondering how long will it be 1st place Quote
gradus Posted September 29, 2015 Posted September 29, 2015 (edited) How you set 66:100 on 1133? Its not available in bios on me st6, only from jumpers. I think its some strange bug 99%, setfsb show 188fsb/141mem/47pci. If use 2:3 then must be 94pci and that impos for FX5500 and IDE hdd. Edited September 29, 2015 by gradus Quote
ludek Posted September 29, 2015 Author Posted September 29, 2015 I wanted to do it again, I can not see that PCI /2 thing. What did I do when ocing?! Maybe U have right with some bug. But I even recorded SPI 1 M Video... I think PCI is limitting me. Tried with MX440-8x card but goes upto AGPCLK>70 MHz or somewhere like that. Quote
ludek Posted September 29, 2015 Author Posted September 29, 2015 (edited) - spi on SDRAM 255MHz. What is going on? I can't make it again... I tried even record a new vid. @gradus try ST6 9P bios. I can send you that ver 2002/07/03 (yyyy/mm/dd). Works normally with 66:100 but I still can't do this again. Edited September 29, 2015 by ludek Quote
gradus Posted September 29, 2015 Posted September 29, 2015 Now try boot 133:100 and run superpi on same fsb with same clocks and memset settings. Imo result will be same. 2.30 its very slow for 255) Quote
ludek Posted September 30, 2015 Author Posted September 30, 2015 That was my first idea when you expressed your doubts. I wanted to run SPI at the same FSB settings and see what will happen. Ok here are some screenshots: 1. The screenshot from the video 2. The screenshot 133:100 a the same settings heck, seems like a bug How else can I explain that? But it means that CPUz has a strange algorithm(?) Well newest version has number of improvements like XOC (and useless bench option) but eg. is saving screenshots at wrong CPU-Frequency-values.BMP when pressing F5. I made some calculations and if the CPU frequency is for example 1653 MHz, filename will be cpu-526.bmp. 1653/526=PI... what the?! I don't blame anybody, maybe it's a bug. I will delete this submission today if after more tests I find it buggy or wrong. Gradus I can send you newest ABIT bioses. yesterday I found a good way to get those from internet. Quote
TerraRaptor Posted September 30, 2015 Posted September 30, 2015 I guess it is a bug)) My good tusl2 stops at 142 with 2:3:1 with vnb 2.27 Quote
ludek Posted September 30, 2015 Author Posted September 30, 2015 I can't explain this. Seems like motherboard turned on after 2:3 (FSB:MEM) set. Then was overclocking flawesly and I thaught that I'll record a videos. After this I wasn't able to reproduce the score because of motherboard decided that exacly the same settings will now work as 4:3 (FSB:MEM). How... But anyway I learned something I have never seen before. Sorry for the mess. I have new conclusions, correct me if I'm wrong: 1. Cache ECC cheching is giving absoultely nothing and might be usefull when long SPI32 run 2. Q.deph 1 is slow 4 is fast (I never was looking at this option thinking it's something not with performance) 3. enabling every option in memset on solano platform and 2-2-2-5 and 15us refresh is the fastest setting 4. good mems is the fundamental thing! 5. Hard disks are crashing at >60-70 MHz PCI clock and some AGP cards are strong *6. It might be need to solder a CMOS divider to divide PCI and AGP clocks... Hmm clock integrity... Maybe it will work 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.