[Sammelthread] AMD K7 - Sockel A (462)

[Sparksnl]

I might have found something: I was using my PCI HD4350 (I have 2) and one of them has issues on some boards. I had the same issue on my NF7. I swapped it for the other HD4350 and that one works fine.
It is new to me it also has issues on Epox. So I might want to retest it later.

Edit: And it works!!!
Edit 2: Too bad the SSD is broken… I suspected this disk already because on Z97 it had issues as well.

 

[R-998]

Greetings to all, and especially @Tzk and @digitalbath , could you tell me for what reasons the bios you modified (I tried EBED of all versions for both boards, since I need a romsip from it to achieve maximum fsb) for a7n8x-e deluxe and a7n8x-deluxe rev2 causes hang on boot after "Verifying dmi pool data...
update success" on the Crossflashed a7n8x-x, the bios from these two boards in the version from Trats work fine on my board, the two-channel works correctly. Could you explain why these modbioses freeze and how to fix this? Or just put the "Hynix romsips" from EBED into Trats modbios for a7n8x-e . it s all i need from mod

 

[Tzk]

Maybe it's the APIC (not ACPI) setting? Usually it tries to load the OS after "verifying dmi pool data". I can confirm that all of these bios mods work and are 32M stable on win XP.

on the Crossflashed a7n8x-x, the bios from these two boards in the version from Trats work fine on my board, the two-channel works correctly.

I've also noticed that this works, but i'm not sure crossflashing is enough to actually enable dualchannel on A7N8X-X. Yes, it reports DC on boot, but is it actually working? You'd have to run the ram async to FSB (ram slower than fsb like 6:4 or 2:1) to doublecheck. That's a test which is still on my list.

 

[R-998]

Maybe it's the APIC (not ACPI) setting?

Nope , i changed , and reset esdc/configuration data too) All peripherals (network, sound, ports) are also disabled. 6:4 works fine on Trats bios , Vnb 1.57v.

 

[digitalbath]

Did you change a setting in one of my mod BIOSes? My guess is that a non workable setting is causing the freeze. But I could be wrong.
I would try one of the newer BIOSes (named with a B2 or a B3) like the EB-3 mod BIOS. The older BIOSes should also work, but the romsip preset (with memory names like Hynix) didn't work as intended. As far as I remember the preset "Winbond" is one of the few that might help.
Try to use higher alpha timings for a higher FSB frequency.

E. I am currently not at home. So I am not able to mod a BIOS for you now.

 

[R-998]

Did you change a setting in one of my mod BIOSes?

Yea , i try any romsips and 2 different start fsb (133 and 166) . alpha timings setted as 6-6-6 , all peripherals are off. On original X-E all works fine with Hynix sips (i have both boards). B2A tested too and has no difference. On original X-E all works fine , on crossflashed X-X - freeze , single-channel and 6:4 tried too

 

[digitalbath]

Do all additional settings at "Auto" work? I want figure out if the option Rom module in the BIOS is causing the freeze. The option Rom module has the code for the "new" BIOS settings. This module is the last module that the BIOS loads just before the system starts to boot from the os.

 

[R-998]

Do all additional settings at "Auto" work?

No , i try auto too , won t boot on crossflashed.

 

[digitalbath]

So, the option Rom is not the problem here. I guess the problem is caused by the Bios version. I used the latest official Bios (or beta) version for the mods. As far as I remember, trats mod BIOS uses an older official version. Maybe ASUS changed something, that causes the error after crossflash?
Maybe worth to try trats BIOS and change the romsip values in Windows with a tool?

 

[R-998]

Maybe worth to try trats BIOS and change the romsip values in Windows with a tool?

Yea , It s works , i save profile with NF2tweaker on X-E with EBED B2A and load on cf X-X Trats , but does it give similar possibilities as EBED modbios?

 

[digitalbath]

The romsip tables in the NF2 BIOSes consist of two parts. The first part is IMHO mainly for the NF2 / memory Controller and the second part lists the CPU FSB timings for every Multi, aka SIP timings.
In the tweaker you see only the values of the first part. We are able to change this values onthefly. So the changes will work.
The SIP values from the second part are loaded during system start. We can not change them after the start, but since we know where they are stored in the BIOS, we can mod the BIOS with the SIP values we want.

The EBED romsips do have the "EB" values in the multi tables (second part). They're slower then the tables from trats BIOS. I am not sure how far the slower values will help in OC. In my opinion, the first part of romsip is more important for FSB OC.

 

[Tzk]

on crossflashed X-X - freeze , single-channel and 6:4 tried too

That's interesting. I've tried the crossflash and it worked flawlessly, even for booting win XP... Else i wouldn't recommend trying it. Besides that, if you set all "new" settings to auto, then nothing should be set/changed by the modded option rom. So if the boot fails, then either your -X board doesn't like the sips or something else has gone wrong.

This issue leaves me a bit clueless, especially since Trats sips usually stress the chipset quite a lot, but EB/ED/EBED are a bit more relaxed. Trats bios is great if you keep the fsb below ~240Mhz. Above that it's ED which does a great job.

 

[R-998]

I've tried the crossflash and it worked flawlessly, even for booting win XP... Else i wouldn't recommend trying it.

But on your EVO modbios for X-Deluxe 2.0 you write "DO NOT FLASH ON non-Deluxe or v1.0 BOARDS" , stock bios for x-deluxe 2.0 are compatible too with X-X (i tried) , what is changed in this modbios that makes it not compatible with other versions of boards?

 

[Tzk]

The EVO bios was made before i knew that it's possible to crossflash and i didn't update the description of that bios.

Another thought... Usually the -X boards got a 256kb bios chip. However the v2.0 and -E boards got 512kb. Have you swapped the chip for a bigger one? I'm not sure the regular bios fits into a 256kb -X chip, even if you force-flash the board. Also i didn't get the -E Bios to work on -X. Only the normal v2.0 bios worked flawlessly on -X.

 

[R-998]

Usually the -X boards got a 256kb bios chip

Of course i use 4mbit chip , sst49lf004a. Latest version of EVO bios cause freeze , but if i use old ver where modded only romsips table (1008-RS-bpl319-ebed-romsip-1T-DLX-v2: , finded on HWbot forum) all works fine , i think custom option rom cause this freezes

 

[Tzk]

Yes, that might be possible. Still if you choose "auto" for all new options, then it should completely bypass the option rom. it basically checks if the options are set to zero (that corresponds to auto in bios) and if yes, then it skips the code segments. Besides that, i'd prefer the newer mods by digitalbath over the old RS bios.

Just to doublecheck: you've cleared CMOS via jumper after flashing, right?

 

[P4schissel]

Gibt es eigentlich auch ein Modbios für das Asus A7V333?

Ich bekomme einen XP 3000+ (200MHz FSB) nicht richtig bei 166MHz FSB zum Laufen (stürzt beim Laden von Windows ab). Bei 133MHz FSB läuft der Prozessor und auch in einem A7N8X Problemlos bei 200MHz.

Auf dem A7V333 läuft eine native 166MHz CPU (XP 2500) ohne Probleme. Meine Vermutung sind die Romsips

 

[R-998]

Just to doublecheck: you've cleared CMOS via jumper after flashing, right?

Yea , i just turn PSU off , flashing use hot-swap metod with AWDflash 8.24F with /cd /cp /py /sn /f keys on original X-E mobo , booting system to dos with trats 1013 bios. I l try all a7n8x series bioses on X-X , all work fine , but X-E needed to be disable halt on errors to No errors , bios from X-Deluxe 2.0 works fine but a little longer post (different sil sata bios ?)

 

[digitalbath]

@R-998 Yes, the mod BIOSes got a new SIL SATA Firmware. It causes the longer boot process.

@P4schissel ich kenne kein mod BIOS für das A7V333. Ich kenne auch keine Tests oder andere mod BIOSe für den KT333 Chipsatz, wo speziell die romsips verändert wurden.
Ich habe mal bei einem KT600 Board an den romsips rum getestet und keinen wirklichen Unterschied bemerkt. Romsips und VIA Chipsatz steht noch auf der to do Liste.
Ich denke das Board sollte auch ohne romsips mit dem 3000+ klar kommen. Bei meinem KT3 Ultra2 hatte ich mit einem 2800+ keine Probleme. Auch nicht mit 185MHz FSB.
Könnte es das Netzteil sein? Vielleicht zu schwach auf der 5v Schiene?

 

[P4schissel]

Das Netzteil ist nicht unbedingt das stärkste, aber bei niedrigerem Multiplikator sollte die Leistungsaufnahme ja passen.

Ich werde das Board erstmal recappen und mit einem anderen Netzteil testen.

 

[Tzk]

Aus der Erfahrung mit dem A7N8X kann ich sagen, dass 15A auf 5V nicht für einen Barton mit >1.8ghz bei 1.65V und mit einer GF4 Ti reichen. Mal so als Richtwert...

(different sil sata bios ?)

Yes. Usually Asus ships the bios with iirc 4.2.50 and we upgraded to 4.4.0.1 or whatever the latest bios for the 3112 is.

Interesting that your -E bios works on -X. Or i misunderstood your post. Also worrying that the boards refuses to boot. I'm out of ideas why that happens, sorry. My -X works perfectly fine with a deluxe v2 bios.

 

[P4schissel]

Das Seasonic kann 20A, aber nur 120W combined. Funktioniert gut mit dem A7N8X / 3000+ / 9800 XXL.

Habe jetzt aber noch ein Enermax EG465 bestellt.

 

[digitalbath]

@P4schissel
Das stimmt wohl. Mit einem kleinen Multi sollte sich die Leistungsaufnahme im Rahmen halten.
Ein recap könnte durchaus helfen.
Wenn ich wieder zu Hause bin, kann ich mir auch die romsip Tabellen ansehen. Vielleicht liegt es doch daran? Voreilig ausschließen möchte ich dann doch nicht. Wäre aber für mich ungewöhnlich.
Man könnte das durch einen Test dann ausschließen. Erinnere mich daran, falls ich das vergessen sollte.

 

[R-998]

Interesting that your -E bios works on -X. Or i misunderstood your post

Yea , works but needed turn off "halt on errors" to boot. Bios from X-Deluxe v2 works fine on my X-X , but only old version without "option rom mods" , all STOCK bioses from A7N8X series works fine on X-X , but all biosmods with changed option rom cause freeze. Chipset is Nforce 2 400 GBB479.1 0415A1 A Taiwan , South - Nforce 2 MCP A4 , A7N8X-X board rev 2.0. On 1.81v vnb/3.05v Vmem (hynix BT 2x256) passed SPI1M at 262mhz on TaiPan {EB}{ED} romsips (my ram limits me i think)

 

[Tzk]

I'll also check what you reported when i got a bit of spare time. Regarding the FSB, try setting the timings to 3-4-4-8 or 2.5-4-4-8 for BT-D43. These need very relaxed timings to reach high clocks. 262 at 3-3-3 is quite good for these chips. Also keep the FSB at or below 263 in bios and use clockgen to reach 265+... At least my boards refuse to boot reliably at 264 and above. I usually boot at 250-255.

 

[bschicht86]

Wenn ich wieder zu Hause bin, kann ich mir auch die romsip Tabellen ansehen.

War da nicht was, dass je nach Standard-FSB an den CPU L-Brücken je was anderes geladen wird? Evtl. kann das Board mit der 200er Einstellung nichts anfangen und lädt z.B. was für 100MHz FSB.

 

[Tzk]

War da nicht was, dass je nach Standard-FSB an den CPU L-Brücken je was anderes geladen wird?

Korrekt. 133 und 166 brücken jeweils einen Kontakt, 200 brückt beide. Also quasi so:

133 = open close
166 = close open
200 = close close

100 müsste open open sein, wobei ich mir da nicht 100% sicher bin. Das ist auch der Grund weshalb man aus einer 133er und 166er Cpu jeweils eine 200er machen kann, 133 -> 166 geht aber nicht. Generell kann ein Mainboard diese Brücken auslesen, muss aber nicht.

Fab51 hatte dazu super Infos... https://fab51.biosbude.de/cpu/barton/athlon-e23.html#L12

 

[R-998]

Also keep the FSB at or below 263 in bios and use clockgen to reach 265+

I booting at 230fsb and up to 260+ using tweaker , 265 and 268 works but memory unstable (spi1m possible a bit randomly , hwbot prime 32m stable , 270+ cause freeze) , timings 3-4-4-11 tried but not help , much Vmem too (3.3v)

 

[P4schissel]

Das A7V333 hat einen ROMSIP jumper. Mit der Einstellung auf "Hardware" läuft nur der 2500+ auf 166MHz FSB. Auf "Software" funktioniert der 166er Barton nicht (166MHz im Bios eingestellt).

Bei dem 200MHz 3000+ funktionieren beide Einstellungen nicht.

 

[Tzk]

Beim 3000+ würde ich erwarten, dass „Hardware“ nicht funktioniert. Weil das Board vermutlichmit dem 200er FSB nichts anfangen kann. „Software“ dagegen sollte klappen. So oder so hat die CPU einen höheren FSB als Unterstützt, da treten mitunter interessante Effekte auf.