| Thread Tools |
20th November 2008, 22:45 | #21 |
[M] Reviewer Join Date: Nov 2004
Posts: 4,127
| Yes I got paint, that is not the issue. The system freezes when I press print screen 90% of the time. It often freezes when I open paint, open task scheduler, ... You see, even taking a photo is a hell of a job, when I try to grab and drag a window, system freezes. For one stupid superpi score it happened that I had to reboot 10 times before I could have every application nicely line-up in order to be hwbot valid. For every benchmark I have cpu-z, memset, clockgen running in the background, times are still okay since I could easily beat the competition cpc when I was testing the 1.5GHz willamette. Yes, better is possible, wprime and superpi 32m, but it takes so long that I'm not really interested to tweak it down just because of the printscreen issue. The setup can run 2h wprime stable, it can run through memtest at high frequency's no problem, but just don't do anything complex like dragging windows At optimal settings the OS uses not more then 30MB memory space though, don't think I'll gain much be using other OS. Or is there more then meets the eye? Last edited by geoffrey : 20th November 2008 at 22:48. |
20th November 2008, 22:52 | #22 |
Madshrimp Join Date: May 2002 Location: 7090/Belgium
Posts: 79,021
| much more; if Massman is to be believed, some copy wazaa trickery can lower your SUperPi times
__________________ |
20th November 2008, 23:02 | #23 |
[M] Reviewer Join Date: Nov 2004
Posts: 4,127
| ah yes, an OCX member has made a small utility which simulates copy waza. I'm not using it at the moment, not needed in order to take first spots in superpi But... it wouldn't heard to have a copy of such well tweaked OS, safes me the time I would otherwise spend on shutting down everything which is not needed |
20th November 2008, 23:44 | #24 |
[M] Reviewer Join Date: Nov 2004 Location: Waregem
Posts: 6,466
| It's not the amount of idle memory usage which is the key for having a good benchmarking os, it's the efficiency of the memory and resources usage that's important. Copy-waza basicly prepares the system for big calculations, which has quite a decent influence on the system's performance. For instance, iirc, the AXP setup would apparenlty have been 0.2s faster in 1M if we'd done a copy-waza. |
21st November 2008, 08:56 | #25 |
Member Join Date: Apr 2005
Posts: 1,338
| server 2003 is best for P4 G
__________________ http://www.hwbot.org/signature/u5776.png |
22nd November 2008, 11:00 | #26 | |
[M] Reviewer Join Date: Nov 2004
Posts: 4,127
| Faster then Superpi OS? Quote:
You see, this board can go quite high on FSB, but with multi 12 I aint going to get much higher in frequency's when using LN2, need to get the multi 14 working. 14x190=4060MHz, now that's something, CPU will probable stop clocking earlier though, 3,8GHz~3,9GHz at most. Its a step B, not same quality as Northwood step C. Would yield in roughly 180% oc, not that bad EDIT: how exactly did you do the CPU microcode bios modification on your PT800 board? Last edited by geoffrey : 22nd November 2008 at 11:04. | |
22nd November 2008, 13:17 | #27 |
[M] Reviewer Join Date: Nov 2004 Location: Waregem
Posts: 6,466
| I'm in doubt about the OS; it might even be so that win2k is faster on this old technology. For the Athlon XP, Win2k is apparently faster. Yeah, I was going to check efficiency, but haven't come across anything interesting so far. It might be cool to find out in how far memory timings and dividers play a role, but that's only interesting for data analysis, not really for breaking records. Is the multiplier of that cpu unlocked? Why doesn't 14x work? I changed the CPU Microcode using bios editing programs, as far as I can remember. Firstly, I extracted the microcode out of the bios of a recent motherboard, then added the microcode to the new bios. It was a fairly simple program to use: completely GUI. I'll search for the program and link you |
22nd November 2008, 13:50 | #28 |
[M] Reviewer Join Date: Nov 2004
Posts: 4,127
| No idea why it doesn't work, 14 is full speed multiplier, 12x is speedstep multiplier. Probable same reason as with the E8600 on PT800, not completely supported. |
22nd November 2008, 14:03 | #29 |
[M] Reviewer Join Date: Nov 2004 Location: Waregem
Posts: 6,466
| What bios are you using? |
22nd November 2008, 15:41 | #30 |
[M] Reviewer Join Date: Nov 2004
Posts: 4,127
| Latest, F23: P4C800-E Deluxe BIOS 1023 This BIOS support mobile CPU adapter card CT-479. Please do read the description before updating BIOS Do not use old version Afudos.exe to flash BIOS.Must use new Afudos(2.07) or later revision to flash BIOS. Using ASUSLiveUpdate to flash BIOS, you need to upgrade ASUSLiveupdate to v5.22.02 or later before flash BIOS. This BIOS is for motherboards purchased in a retail ASUS package. If your motherboard is part of Compaq X07 Gaming PC or Compaq X09 Gaming PC systems, you should obtain BIOS updates from HP, as using this BIOS may have adverse effects, such as rendering your system unusable. Support CPU multiplier item when using Dothan 533 with CT-479 F24 is beta BIOS, don't see why that BIOS would suddenly support the Pentium 4 mobile since these CPU's were launched way before BIOS F24 (y: 2005) |
Thread Tools | |
| |