[Sammelthread] AMD K7 - Sockel A (462)

[The Sandman]

Hmm... leave all the RAM settings on optimal and set the NB Voltage to 1.9. Maybe it works

 

[Sparksnl]

I think I just succesfully booted the tmod cd from usb.
All ram settings were on auto once. I will look into it later.

 

[Tzk]

iirc i burned the tmod cd v11.1.1 onto a regular 800mb CD. I've never used the v11.2.1 as i sold the dfi board before i found a working link online. Nontheless you should be able to boot it from USB or even in a VM.

 

[Sparksnl]

I have a working usb boot now:



I found the bios files on the disk and I can access them. I just needed the menu. How it looked like.

 

[The Sandman]

You can also extract the ISO using WinRAR. The .ima files can be opened and extracted with Winimage

Download WinImage

 

[Sparksnl]

I allready extracted it a while ago and noticed that the .IMA files are zip files. But thanks.
The .BMP files are pictures and the EZB files are partially readable as txt.
I used GRUB to make a bootable USB.

Edit: Strange that you get more then the .bin file. I only see it's bin file in LPB619XT.IMA
Edit 2: Ah I see WinImage opens more.

 

[The Sandman]

WinRar is also showing only a "original.bin" file. 7zip shows all files.

 

[Sparksnl]

Thanks! I was wondering if I should try 7zip. I was just one step behind.
I was wondering if other parts are Linux based.

 

[The Sandman]

The CD is Linux based. You can find Prime95 in the tools section. There is no Prime95 for DOS^^
There is an even more advanced cd floating around but ment for newer systems (Athlon XP and older stuff doesn't work):

mprime multithread live CD

EDIT: 13.10.2013 Update neue Version mit lm-sensors und root autologin LiLi. 13.10.2013 Update GNU screen eingebaut für bessere Übersicht THX COLOSSUS. 28.10.2013 Download link geändert. 01.12.2013 Neue Version mit kernel 3.10. Habe bemerkt dass bei der alten Memtest nicht funktioniert, wurde...

www.overclockers.at

Kann das mal jemand von euch LP B'ern mit den 270-MHz nachstellen? Memtest als Screenshot reicht. Kein modifiziertes Bios nötig...
bzw. was brauchts an NB Spannung für 263-MHz bis in's Windows und Valid?

 

[Sparksnl]

That explains why I could not “unzip” other files. Those are probably “zipped” in a Linux format. The older files were Winzip because it was the best programming for compressing. The fact that he stated that he had to use Winzip gave it away.
I know what to do then.

 

[Strunkenbold]

@Strunkenbold
Kann es sein das das Asus A7N8X-E doppelt in der HWBot Datenbank ist? Einmal als A7N8X-E und einmal als A7N8X-E Deluxe? Ein -E non-Deluxe gab es meines wissens nicht?!

Anhang anzeigen 537631

Danke für den Hinweis! Alles was ich an subs finden konnte, wurde zum Deluxe geschoben. Leider sind 2 subs auch für mich als Admin nicht sichtbar, so dass die zwei Kategorien erstmal weiter bestehen werden. Das muss sich ein Programmierer ansehen...

 

[The Sandman]

Chef DB8 solls richten
verrückte Zeiten...

 

[jumpel]

Endlich auch ein paar Photos von mir:
Abit AN7 - V1.0 - V0.32
NB: - SB: - back: - VCore :]

 

[stunned_guy]

UPDATE und Fehlerbereinigung.

Die Seriennummer kannst du bitte noch mit angeben, auf dem letzten Bild ist das die falsche Nummer. Danke

 

[The Sandman]

@Strunkenbold das hier ist auch falsch drin:

UE50`s wPrime - 32m score: 54sec 593ms with a Athlon 64 3500+ (Newcastle)

The Athlon 64 3500+ (Newcastle) @ 2904MHzscores getScoreFormatted in the wPrime - 32m benchmark. UE50ranks #null worldwide and #5 in the hardware class. Find out more at HWBOT.

hwbot.org

steht beim Ultra B https://hwbot.org/hardware/motherboard/lanparty_nfii_ultra_b/

 

[jumpel]

Die Seriennummer kannst du bitte noch mit angeben, auf dem letzten Bild ist das die falsche Nummer. Danke

Die auf dem weißen Aufkleber meinst du? Habe mich an deinem Post 3096 orientiert, da war auch keine solche Nummer ;]
Hier bitte:
RD460408028019957 ... das vorletzte müsste eine 5 sein.


Und weiter geht's:


Abit NF7 "non S" V2.0, schwarze RAM Slots, Nr.: RA890330040042459
Board: - NB: - SB: - Nr:

Beitrag automatisch zusammengeführt: 10.11.2020

Shuttle FN41G/FN45 V1.0, Nr.: FN450000002292
Board: - NB & SB:

 

[Sparksnl]

New arrival: IQYFA 0343 mobile on nf7-s


Looking good so far...Quicktest, boot only. Not much time.
2600MHz needed something like 1.725-1.750v.
It will probably run better on my Epox.

Edit: I just quick tested on my Epox as well.
2500MHz: 1.600v
2604MHz: 1.700v
2695MHz: 1.800v
2784MHz: 1.975v

I did a boot only. BIOS voltages.
Not good enough... by 0.25v for nr2.

 

[jumpel]

Whoa, nice results!
That's a good chip. I think my best one needs 1,825 V for 2,6 GHz (BIOS, NF7-S) to run 2h in prime.
It's a big gap to 1,675 at 2,5 GHz :[

 

[Tzk]

Note that his results are for boot only, no 32M or prime. It's enough to pre-bin fast though. My chip needs a bit more voltage than his, but for prime stability 1.825V for 2.6ghz is an above avarage (aka decent) chip, but not a golden sample. I'd be happy with it nowadays.

 

[ItsFun]

I have two IQYFA 0347 VPMW and a IQYFA 0349 XPAW. None of these cpus runs better than 2400 MHz @ minimum 1,7V. But i only testet on Shuttle SN45Gv3.

 

[Sparksnl]

Yes that is correct. I would have to run PI32M or prime it to know the exact results for stability.
I might run PI32M but I doubt I will be priming it since I don’t want to put too much load on these cpu’s.

Edit: There is always a bit luck of the draw necessary...

 

[Tzk]

Yes, for benching purposes i'd test it with 32m. No need to torture it for hours with prime.

@ItsFun
I got a 2600+ AQZFA 0347 VPMW from bschicht on the Verschenkethread. We'll see how good it performs. As it's not a mobile it'll be superlocked. My 0406 UPMW pass 2.4ghz at 1.65V, the best needs 1.7V for 32M at 2.5 and the other won't pass 32M at 1.75V. Third one is dead. So yes, you need a good stepping and luck as the differences between chips seem to be fairly big.

 

[WMDK]

IQYFA

There's your explanation already. That stepping, although being an "I" stepping isn't known for being a good overclocker since the "F" tells us it's still based on the earlier sub-revision before AMD made some fundamental changes to the process. Achieving good results with steppings based on early revisions can be quite excruciating. Therefore your results are within what should be expected. What you really want are the steppings which were designed specifically with the new process in mind. From the "worst" to the best these are: IQZFA, IQYHA, IDYHA. IQYHA in fact was THE top performer up until 0447/0448. After that, IDYHA was introduced and manufactured from at least 0449 until almost mid 2005 and was even a tad better than IQYHA CPUs. All these three steppings came with their less perfect siblings, the AQZFA, AQYHA and ADYHA. In late 2004 / early 2005, when AMD had fine tuned the process to the max, "I" and "A" steppings performed almost identical. However by that time, almost everyone was already captivated by the Ahtlon 64. Mobile XP CPUs of that time are hard to acquire since most retro hardware fans know about their value and capabilities which is why usually, you'll only find the dull rejects from early to late 2003 on eBay and IF there's a good one for sale, there's almost always a gleaming price tag to it

 

[The Sandman]

everytime he does this, I feel dull....

 

[ItsFun]

There's your explanation already. That stepping, although being an "I" stepping isn't known for being a good overclocker since the "F" tells us it's still based on the earlier sub-revision before AMD made some fundamental changes to the process.

I know that. Thats why i am surprised of the results from sparksnl.

 

[Sparksnl]

I tried some idyha in the past but I did not have luck with these chips yet.

edit: I found some good results here:

WHAT IS A MOBILE BARTON?

Madbox nos entrega una pequeña descripción acerca de los procesadores mobile

www.madboxpc.com

edit 2: copy/paste from above link. Seemed good enough for me.

IQYFA 0343 MPMW
(Water, DipStickTony) (Chevvy core, Neomoses Nemesis)

Before burn in:
* 2708 MHz at 1.8 volts. Load 45C
* 2750 MHz at 1.85 volts
2.1 volts on NF7-S will not boot into Windows at 2.8 ghz

After burn in:
* 2700 MHz at 1.65 volts (1.7 in the BIOS, board undervolts Vcore). Load 41C
* 2800 MHz at 1.875 volts (1.95 in the BIOS, board undervolts Vcore). Load ~46C
2950 MHz at 1.95-2.03 volts (2.2 in the BIOS, board undervolts/fluctuates Vcore).
2950 MHz Runs Prime95 Priority Ten for one hour, then crashes.

 

[WMDK]

I know that. Thats why i am surprised of the results from sparksnl.

A golden cpu really. You don't see that on a regular basis.

 

[Tzk]

* 2700 MHz at 1.65 volts (1.7 in the BIOS, board undervolts Vcore). Load 41C

I seriously doubt these "after burn in" results. I've never seen a cpu which did 2700 on ambient cooling and below 1.75V. I even remember that a few guys on HwLuxx managed to squeeze XP-m to 2800 at 1.9xV, there's just no way those results are anywhere near stable. The "before burn-in" results look valid, that's what my cpu can also achieve.

 

[Sparksnl]

I am not sure about the after burn in results as well. But these and others were good enough to go for it.

 

[Tzk]

Yes, any cpu which manages a 32M run at 2500Mhz and 1.65V should be considered "very good" and if it passes at 1.625V or lower it's probably a golden sample.