Jump to content
HWBOT Community Forums

Tzk

Members
  • Posts

    229
  • Joined

  • Last visited

  • Days Won

    9

Everything posted by Tzk

  1. VTT is generated from Vdimm (yes, seriously!) and thus is tracking Vdimm. You can even use the boardviewer archive terraraptor linked above to get the pinout of every single component on the A7N8X. Tracing Vtt is very easy then To measure Vdimm i suggest to use the point on the VTT mosfet, right next to the two caps. That's "after" the copper coil. The other measurement point (bottom of picture) ist "before" vdimm passes the coil and thus about 0,1V higher. I tried tracing down VTT recently, so here you go:
  2. You can try that, but expect serious issues when reaching 245 to 250Mhz. At least with 512mb Kits. I currently got 6x 512mb TCCDs (4x Corsair XL, 2x Geil Ultra-X) and only managed to get one of the Geil sticks through 32M in singlechannel. I'm not sure what's holding me back right now, as all other sticks fail, regardless of Vdimm (tried 2.6 to 2.8V). ED55 Bios is fine, as long as you got all "extra" options like Drive strength, Slewrate and Alphatimings. So far Drivestrength 4 and Slewrate 10 works best on my TCCDs. I also found that settings Trc to 14 (instead of default 13) and Trfc to 16 (default 15) and using CL3-3-3-11 seems to improve stability above 240Mhz. Tref of 2560 seems to also work well. FYI ED55 are the romsips made by Infrared for his NF7 and AN7. We ported those as they are similar to EBED/ED but seem work a tiny bit better. And last but not least: I found Hynix 256mb Sticks work very well on my A7N8X. Was able to reach 264Mhz FSB with just 2.7V. My board is seriously modded though. Vdd, 12V rail mod, Vdimm, extra caps on vdimm.
  3. If you got an -E Deluxe, then @digitalbath made a "new" modbios, if it's a regular Deluxe v2.0 then i can upload my port of it. Both got the Alphatimings, Drivestrength, Slewrate and Romsip options directly in bios. We noticed that some board won't respond to Vdd and some take a lot of it. So make sure to check if your board likes high Vdd. He got a board which runs 250Mhz+ on 1.6V while my A7N8X Dlx needs 1.85V to reach 260Mhz. and a question for @I.nfraR.ed: What does the "drive strength" option in your tweaker do? I can choose "normal" and "weak", right? Does that mean that choosing DS 1 to 15 will result in 15 "normal" and 15 "weak" options? Or are these 15 options partially normal and partially weak, just like on DFI NF4? (DFI got "strong DS" for the even options: 2/4/6/8 and "weak DS" for the uneven options 1/3/5/7). Currently it seems like all ram chips i tested liked DS around 3 to 4, even winbond. i haven't tested the "weak" option though.
  4. Sure! It refers to the romsips inside the bios. Ed and ebed is taken from merlins taipan bios and those are based on dfi beta bios. Dded romsips are made by digitalbath. iirc ed is fastest of the three, but gives you the lowest fsb. I found that Ebed gives a bit more fsb but the tradeoff is a bit less efficiency. Dded is again a bit slower than ebed, but perfect for max fsb runs. I‘m able to run 32m at 263mhz fsb with ebed romsips (see above), i‘d say try for yourself, but ed or ebed isn‘t a bad choice
  5. I'm still trying to find the FSB limit of my boards chipset. @digitalbath suggested to try Hynix BT-D43 or DT-D43 on 256mb Sticks. Not efficient, but seems to run well on nforce 2. So, here we go 260Mhz 32M... That's the first try with those Hynix. Seems like my Infineon 6A are currently limiting me at 257Mhz 2x256mb Hynix BT-D43 | 2.9V | EBED Interface On | BPL 3.19 1T 260Mhz x 9 | 3-4-4-11 A7N8X Deluxe v2.0 | Vdd 1.85V Aircooled: Zalman 7000Cu on XP-m 2600+ and stock NB heatsink + 120mm fan blowing onto it. EDIT: 263Mhz passed 32M. However i can't complete a 264Mhz run, even when i raise the Vdd from 1.85V to 1.925V. I guess that's the limit of the board. Got a blackscreen after 10 loops.
  6. You gotta be kidding me... ? How awesome is that... I'm out of words right now. That's a powerful tool!
  7. I've bought some more ram and tried to get 2x512mb stable on my a7n8x. Sadly it seems to be impossible to get TCCDs stable above 240MHz. I've now got 4x512mb Corsair 3200XLs (TCCD), 2x 512mb Geil Ultra-X (TCCD) and 2x512mb Corsair 3200C2 v4.2 (TCC5) and all Sticks run into a wall at 240Mhz, regardless of voltage or timings. I even flashed the SPD on the 3200XL to disable the 2nd rank, resulting in a 256mb SS stick. But that changed nothing, the 240Mhz wall is still present. However my Infineon 6A (256mb) can run up to 257Mhz 32M at 3.6V flawlessly. Question is if the Vdimm VRM is at its limit or if the NF2 can't handle doublerank sticks. I noticed that with 2x256mb Winbond, the Vdimm rises about 0.1V when under load (3.6 ->3.7V). Have you had issues with fluctuating Vdimm on the A7N8X @Terraraptor ? --- Now two thoughts: a) why is the OCZ DDR Booster listed as incompatible for the A7N8X? Can't we feed external Vdimm? If so: why? b) How on earth is the A7N8X generating VTT? I attached a close-up shot of the Vdimm regulator on the board. The RT9202 generates the VDimm and is the chip where one applies the Vdimm mod (FB pin). but there's no other chip to generate VTT (1/2 vdimm). I booted the board and checked some voltages. Here's what i found: VTT is generated from VDimm through the mosfet next to the two black caps. There's some sort of feedback loop throught the small fet onto the big mosfet (shown with red traces). I've not found the regulator for the gate voltage yet. Not sure if there actually is any regulator?! When running 2.9V Vdimm i measured 2.15V as gate voltage on the big mosfet. I also noticed that Vdimm is generated from the 5V rail. EDIT: So i investigated further and found a small resistor which has the Vdimm on one and the VTT on the other side. It seems to feed the opamp right next to it... The opamp then drives the VTT mosfet. So it looks like the VTT is generated by a simple resistor offsetting the Vdimm and driving the opamp... Interesting! EDIT2: As i'm not an electical engineer... Is this plausible or completely wrong? I got no clue...
  8. I'm currently testing my xp-m with L12 hardmod. Interestingly it performs differently when i set the L12 bridges to 166MHz. Even when the romsips are completely identical... And i don't get why. How big is the usual margin of error on 32M? I got a difference ofseveral seconds on my runs. results: https://www.hardwareluxx.de/community/threads/amd-sockel-a-thread.584473/page-76#post-27462451 I've been able to complete a 32M at 258MHz with Merlin EBED romsips. My board is not able to achieve that with ED romsips though. ED seems a little faster than EBED, but i lose about 3-5Mhz max. fsb. So i guess 253-255Mhz with ED sips is possible on my Asus. No way i could get this running at 260 as that'd require >1.9V VDD and better (probably below ambient) cooling on the nb. Watercooling doesn't seem to be enough for this or i need to lap the nb as it is nowhere near flat. Here's the result: https://hwbot.org/submission/4249058_tzk_superpi___32m_athlon_xp_m_2600_(barton)_31min_45sec_984ms
  9. I believe it was not related to .net or your code. seemed like the download itself (or the extracted files) were broken. i downloaded it another time and now it works flawlessly. @digitalbath started to have a look at efficiency... I now get why you were so concerned about that with the romsip tweaks... Seems like his DDED romsips are very loose, which results in a 10-20sec penalty on 32M. That's a huge loss! We chose 9x240 and a barton cpu as test setup, because every romsip and almost any barton should be able to run at that speed on aircooling. Interestingly my system is a lot faster per clock... (his 32M: 39m 38s, mine: 39m 21s) Have you compared APIC on/off and using the SIL SATA controller vs. a IDE/sata converter on the IDE ports? Is there any difference in performance? I'm running APIC on and IDE hdd. Besides that we're running exact same settings. Both Bartons, both 2x256mb Infineon 6A, same Romsips, same BPL, identical clocks.
  10. I just tried the newest Tweaker and it crashes on WinXP... Does it need any extra library the older build didn't need? 20200416 works, 20200426 doesn't. It shows the splash screen and that's it. Afterwards there's the usual crash screen by windows, nothing else. Nevermind. Download was broken... It works as expected
  11. +1 for the one with tabs. As long as all memory timings are shown on a single page everythings fine. I like it!
  12. Yep. Interestingly i've had CPU-Z v1.91 working on my XP setup before and was able to validate at 250Mhz FSB+. Now it just freezes the system at the Processor detection stage (15% loading), even when underclocked to 166x10. I haven't made any changes to the system in the meantime besides a few reboots. So no clue what happend to cause these issues... v1.59 still works as expected though. My XP isn't tweaked, it's basically factory installed.
  13. I see where your issue lies.. I just tried to use 3 columns, but that just doesn't work out... Here's my paint.exe mockup. I'd rather not do it like this... but the window at least has a bit less height.
  14. As i'm mostly (or only) benching Socket A: Which CPU-Z version is acceptable for you? I've seen v1.59 being mentioned a few times on other sites... I also haven't tried to validate a score with such an old cpu-z version... Will that work?
  15. That last build of NF2 XT works as expected on my A7N8X. You could add the other settings aswell. So DriveStrength, Slewrate, Romsips (only the confirmed ones), Data Scavenged rate etc. I guess that's idea behind the 2nd tab on the tweaker? Also: what about using only a single tab, just like A64 tweaker. That way you don't have to open the tweaker twice to show all settings at once. I've ported @digitalbaths romips tweaks to the non-E Deluxe v2.0 and i'm currently trying to get 2x512mb TCCDs stable above 240Mhz. Let's see if that turns out to be a FSB wall of my board...
  16. That one looks great! What about an option to load and apply the values during windows boot? I love how everyone is still pushing socket a as hard as possible. Thanks for investing your time to create a new tweaker Regarding the romsips: @digitalbath even went further and integrated these in shuttle, nf7 and asus bios. It‘ s all uploaded to the bierbude server, including the option rom sourcecode. Maybe you can take some settings from there? Heres what he did (sorry, german forums), the picture should speak for itself: https://www.hardwareluxx.de/community/threads/amd-sockel-a-thread.584473/page-63#post-27414975 Might also be of interest for @TerraRaptor as these bios seem to run extremely well on asus. bierbude.spdns.org:2302/USER%20UPLOADS/digitalbath/ Sadly i haven’t had time to port his a7n8x-e changes to the non -e dlx yet. He even manged to get 2x1gb running at ddr500 on nf2 ?
  17. I've got several romsip tables but haven't found one which uses your 2nd line. Merlins romips all got 40h at offset 11h and most other romsips i checked have a different 2nd line (offset 10h to 1Fh). DS 4 also works best for me on TCCDs and Infineon BT-5. Haven't tested on Winbonds yet. Also didn't test different SR yet and stuck to the default (10). I'm still trying to get 2x512mb TCCDs stable above 240MHz, but something is holding me back. Also CL2.5 seems to work better than CL3 which doesn't make sense. Right now i'm a bit short on time to work things out, but my plan is to try to combine different romips or (like you did above) to test single settings for bandwidth and stability. There must be a way to get TCCDs stable at or above 250MHz...
  18. You're using custom romsips? Looks interesting... I've forgot to put PLOP into my Asus bios... Might to that too. The Asus is picky about USB booting sometimes. What's left to do now is search for romsips and test them. I extracted some if them from DFI beta bios and noticed that some revisions only got minor differences. So that might be a good starting point. Our only chance is to A/B test different sips and compare bandwidth and stability. A very time consuming task...
  19. I've seen some modbios which mentioned 3.2x, but i still believe someone either mistyped or just changed the version number on a random NVDAMC as this is the only mention of this version.
  20. If you sort them by date, you get this. v4.35 is older than v4.40 which doesn't make sense. Maybe they branched 4.35 off and developed it seperately? At least any 4.x version except 4.35 doesn't seem to work on old boards... NVDAMC/05/12/2003/v3.02 NVDAMC/06/10/2003/v3.04 NVDAMC/08/25/2003/v3.17 NVDAMC/09/08/2003/v3.19 NVMM/4.40/04/29/04 NVMM/4.35/07/02/04 NVMM/4.62/09/06/04 NVMM/4.85/03/07/05
  21. I noticed that the date which is included inside the nvmm doesn't match the version. A higher NVMM version is older than a lower NVMM version - that doesn't make sense. However NVMM v4.35 works on older boards like NF7 and A7N8X. No clue why 4.40 and 4.62 doesn't. Maybe the latter ones need other additional code or newer (hardware) features which the older boards are missing.
  22. And he found another one: GA-7N400S-L, NVMM 4.62. Both 4.40 and 4.62 are untested but i suspect they'll work. nvmm462-cpc-on.bpl
  23. I don't believe this either... Sooo, it looks like there's an even newer BPL available. @digitalbath from hardwareluxx forums found BPL v4.40 inside the NF7-S2G bios. nvmm440-cpc-on.bpl
  24. Yes, i also noticed these changes. We'll have to binary compare to ind this trick. I'm unsure if this trick has any downsides and if no, if we can port it to other boards. I'm curious if it has an impact on the efficiency on SuperPi.
  25. My guess is that the data inbetween the $ml$ and the "bios features setup" string actually does something to the bios items. same for the last few lines of the block, right in front of the next $ml$ string. I haven't succeded in decoding these, though. --- Someone over at hardwareluxx.de forums asked if it's possible to backport BPL v3.19 to K7N2 Delta2, as these boards are shipped with NVMM v4.35 by MSI. Let's see if this helps MSI boards to reach higher clocks... http://bierbude.spdns.org:2302/USER UPLOADS/Tzk/MSI/K7N2 Delta2 Platinum/
×
×
  • Create New...