franck delattre is preparing an universal tweaker
franck delattre is preparing an universal tweaker
I love XS
Road to CompTIA A+ (601, 602) and Network+ (N10-003) Certifications
I want a Penryn 3.33G 12MB & quad sli Nvidia G92
I had another play with it last night, after turning off mem slot 3 and 4 and tweaked a few mem timing, my HTT significantly went up?? Tried to replicate the results without tweaking and it BSOD'ed straight away... and then worked again when I tweaked it again
Gone from unstable @ 300 to stable @ 315.., so anyone else on AN50R might wanna try that
Tried again, same as before. Narrowed it down to TRC, TRFC, or disabling the mem slots.
Before 280 HTT BSOD, now 330 HTT @ default HTT voltage...
What did you do??? Can you paste a screenshot? Amazing increase... I'm stuck at 285HTT...
|| Core 2 Quad QX6850 ES @ 3.5Ghz 1.35V || Thermalright Ultra 120 || Asus P5K3 Deluxe || Gskill F3-12800CL7D-2GBHZ
|| XFX 8800GTX || Dell 2005FPW 20" || Ultra X-Pro 750W LE || 3 x WD 320GB SD + 1 x Hitachi 500GB
Thats great news man, no need to get a new moboOriginally posted by Ragnarok
Tried again, same as before. Narrowed it down to TRC, TRFC, or disabling the mem slots.
Before 280 HTT BSOD, now 330 HTT @ default HTT voltage...
BTW: new version with save/restore and startup settings will be available in about a week.
I can't waitOriginally posted by CodeRed
Thats great news man, no need to get a new mobo
BTW: new version with save/restore and startup settings will be available in about a week.
CodeRed
You're doing an awesome job. Thank you!
i5 750 4.20GHz @ NH-D14 | 8GB | P7P55DLE | 8800U | Indilinx SSD + Samsung F3 | HAF922 + CM750W
Past: Q6600 @ 3.60 E6400 @ 3.60 | E6300 @ 3.40 | O165 @ 2.90 | X2 4400+ @ 2.80 | X2 3800+ @ 2.70 | VE 3200+ @ 2.80 | WI 3200+ @ 2.75 | WI 3000+ no IHS @ 2.72 | TBB 1700+ @ 2.60 | XP-M 2500+ @ 2.63 | NC 2800+ @ 2.40 | AB 1.60GHz @ 2.60Originally Posted by CompGeek
I've just bought 2x256 OCZ PC3500EB... and can't pass 230HTT!Tried Cas 2'5-3-3 and nothing... can't pass that without a BSOD
Can you OCZ EB users post your A64 tweaker values? FSB/values...
I'm really
All voltages at +0'3V, plus Vdimm at 2'9V.
It does 225HTT at cas 2-2-2 estable
At 228 (2'5Ghz) (always with 11X multiplier), it says not exact in round at superpi (at any timing), with 1'8Vcore.. before was stable at 2'7Ghz with 1'75V fsb280... something is going wrong!!! It should be a memory problem, because with an HTT divider I can pass 275HTT.. but memory stills failing at >225
Any ideas? Bad memory? bad luck? Should I try diferents slots?
PD: rig in da firm + win2k3.
Last edited by krampak; 06-29-2004 at 07:42 AM.
|| Core 2 Quad QX6850 ES @ 3.5Ghz 1.35V || Thermalright Ultra 120 || Asus P5K3 Deluxe || Gskill F3-12800CL7D-2GBHZ
|| XFX 8800GTX || Dell 2005FPW 20" || Ultra X-Pro 750W LE || 3 x WD 320GB SD + 1 x Hitachi 500GB
I thought that memory was like the PC3700 one... just read a review from pcstats:
If I've known it before buy them....We started to slowly raise the clock speed of the motherboard a few MHz at a time, at 217 MHz we had to increase the memory voltage a bit to 2.8V, or stock voltage of the memory. To our surprise after that we were able to get the memory all the way up to 230 MHz while using memory timings of 2.5-2-2-5! Loosening the timings did not allow us to get any higher but still this is pretty impressive for high speed DDR!
|| Core 2 Quad QX6850 ES @ 3.5Ghz 1.35V || Thermalright Ultra 120 || Asus P5K3 Deluxe || Gskill F3-12800CL7D-2GBHZ
|| XFX 8800GTX || Dell 2005FPW 20" || Ultra X-Pro 750W LE || 3 x WD 320GB SD + 1 x Hitachi 500GB
Works fine Epox 8HDA3+ (via)
Good news, version 0.2 is now ready for testing
Download: a64tweaker_V02.zip
New Features
[list=1][*]Added save/restore of the register setting to a text file.[*]Added option to apply the settings at startup.
This feature uses the HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\Run registry keys.[*]Added a silent command line mode so that you can set your register settings using desktop links etc etc.
Usage: a64tweaker.exe [filename][/list=1]
Register File Format
The register files are written in the standard windows INI file format. This makes it easy to read and edit by hand. A typical register file is shown below:
Each field is optional, so if you only want to change certain fields (and leave the rest up to the BIOS) then just comment out or delete what you dont want changed. The following is an example of a register file that will only set the Tcl (CAS Latency) value to 2.0. All other parameters will be unchanged.Code:[DRAM] Tcl=1 Trc=1 Trfc=3 Trcd=2 Trrd=2 Tras=5 Trp=2 Twr=0 Twtr=0 Trtw=0 Tref=11 ReadWriteQueueBypass=3 ControllerMode=0 ECCMode=0 BypassMax=7 TwoTTiming=0 MaxAsyncLatency=6 ReadPreamble=4 IdleCycleLimit=3 DynamicIdleCycleCounter=1 MEMCLKFrequency=7 MEMCLK0=1 MEMCLK1=0 MEMCLK2=0 MEMCLK3=0
The values in the register file correspond to the actual bit field values. In the future I will make it simpler so that you can have stuff like the following but I just ran out of time:Code:[DRAM] Tcl=1
Things to doCode:[DRAM] Tcl = 2.0 ; sets CAS Latency to 2.0
[list=1][*]Find out what the reserved CAS latency fields correspond too.[*]Add some error checking and reporting. An error in the register file is silently ignored[*]Learn how to format a post nicely
[/list=1]
Enjoy, and please report any bugs via PM.
Last edited by CodeRed; 07-01-2004 at 02:23 AM.
ThxOriginally posted by CodeRed
Good news, version 0.2 is now ready for testing
Download: a64tweaker_V02.zip
New Features
[list=1][*]Added save/restore of the register setting to a text file.[*]Added option to apply the settings at startup.
This feature uses the HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\Run registry keys.[*]Added a silent command line mode so that you can set your register settings using desktop links etc etc.
Usage: a64tweaker.exe [filename][/list=1]
Register File Format
The register files are written in the standard windows INI file format. This makes it easy to read and edit by hand. A typical register file is shown below:
Each field is optional, so if you only want to change certain fields (and leave the rest up to the BIOS) then just comment out or delete what you dont want changed. The following is an example of a register file that will only set the Tcl (CAS Latency) value to 2.0. All other parameters will be unchanged.Code:[DRAM] Tcl=1 Trc=1 Trfc=3 Trcd=2 Trrd=2 Tras=5 Trp=2 Twr=0 Twtr=0 Trtw=0 Tref=11 ReadWriteQueueBypass=3 ControllerMode=0 ECCMode=0 BypassMax=7 TwoTTiming=0 MaxAsyncLatency=6 ReadPreamble=4 IdleCycleLimit=3 DynamicIdleCycleCounter=1 MEMCLKFrequency=7 MEMCLK0=1 MEMCLK1=0 MEMCLK2=0 MEMCLK3=0
The values in the register file correspond to the actual bit field values. In the future I will make it simpler so that you can have stuff like the following but I just ran out of time:Code:[DRAM] Tcl=1
Things to doCode:[DRAM] Tcl = 2.0 ; sets CAS Latency to 2.0
[list=1][*]Find out what the reserved CAS latency fields correspond too.[*]Add some error checking and reporting. An error in the register file is silently ignored
Enjoy, and please report any bugs via PM.
Road to CompTIA A+ (601, 602) and Network+ (N10-003) Certifications
I want a Penryn 3.33G 12MB & quad sli Nvidia G92
wow amazing!
CodeRed, I love you, will you marry me?
Anyone played with this on the 939 FX's yet? This is absolutely incredible. TY very much.
Now you're freaking me outOriginally posted by HKPolice
CodeRed, I love you, will you marry me?
Umm, what do those mem clk 0 - 3 settings do?!? I'm surprised no one asked yet
They enable/disable the four memory banks supported by the A64.Originally posted by HKPolice
Umm, what do those mem clk 0 - 3 settings do?!? I'm surprised no one asked yet
Most BIOSes will correctly set these values ... except for the AN50R which enables all four memory slots irrespective of how many memory sticks are actually present.
Ragnarok, got a nice speed bump by setting these values correctly (using AN50R).
Cool THANKSOriginally posted by CodeRed
They enable/disable the four memory banks supported by the A64.
Most BIOSes will correctly set these values ... except for the AN50R which enables all four memory slots irrespective of how many memory sticks are actually present.
Ragnarok, got a nice speed bump by setting these values correctly (using AN50R).
Also, would increasing the Tref (using 200Mhz 3.9us instead of 133Mhz 7.8us) help stabilize the ram since it refreshes more often?
What's the difference between a 200Mhz and 133Mhz setting anyways? I thought the 3.9us or 7.8us was the milisecond refresh rate and not the Mhz?
Last edited by HKPolice; 07-01-2004 at 07:52 PM.
Ok I'm having probs every reboot. It says something somethin "cannot connect to device". I think it's cuz windows is still loading all the devices 'n stuff, is there anyway to delay the startup of this prog for 10, 15sec so it loads after the devices are loaded?
I get about 10MB/s more in sandra with dynamic idle timer DISABLEDOriginally posted by CodeRed
max async latency .. the lower the better, but I cant run below 6ns at max FSB
read preamble ... doesnt seem to do much, 4ns is good
idle cycle limit ... 16 clocks (irrelevant with dynamic enabled)
dynamic ... enabled
also set read write queue bypass & bypass max to their max values
Came across this and must say its excellent
I have one small problem, when changing the cas lat setting windows just freezes, I tried 3 sets of ram and still the same, CO stepping cpu.
could someone code one of these for the P4 S478 or does it already work?
Overclocking made in Austria
@HKPolice,
I see a couple of problem with using the sandra mem banchmark as a reference.
Firstly, 10MB/sec is only a small gain and possibly due to the variantion between runs.
Secondly, the sandra mem benchmark has a very sequential memory referencing pattern. It is very localised and greatly benefits from in-page access cycles. By enabling the dynamic idle cycle counter the mem conroller is probably prematurely invalidating pages and thus inccurring a performance hit. If this is truely the case then try disabling the dynamic idle cycle counter (the mem controller will then use a static counter algorithm) and set the Idle Cycle Counter to its max value (256 clocks). This should ensure that mem pages remain valid for as long as possible. Test it and let me know how it goes.
Thirdly, sandra sucks as a benchmarking tool
@the.m,
Most memory with the A64 only works at either CAS2.0 or CAS2.5/3.0. Not many sticks work with both cas2.0 and cas2.5 (XMS3500C2 may be an exception here). My kingston sticks wont run anything but cas2.0 or 1.5. CAS2.5/3.0 just freezes. There may be a combination of parameters that helps out, but I dont know what that is (yet).
@EmineM,
cpulloverclock is working on a universal tweaker, so drop him a line.
amazing work! keep it up!
Posting Permissions
Reply With Quote
Bookmarks