[Sammelthread] Sockel 939 / 940 / 754 - Thread

[Raiden Zero]

Same batch as WR 4000+ by gianni
Mine isn’t super impressing tho. 0631APMW beats it same as 0631GPMW

 

[Sparksnl]

Well I bought it from someone from Italy…
Edit: My KAB1E 0631GPCW did 3300 at 1.55v.

 

[stunned_guy]

0626 FPMW ist eines der besten und letzten Steppings, die überragend gingen. Aber vorsicht: auch hier gab es Ausschuss. Aber 90 % gingen wirklich richtig gut, auch unter Luftkühlung

 

[Sparksnl]

Tested as working. Asus K8N-DL



There is one mistake on the board below. Who is sharp enough to see it?
(I noticed before testing it in a final inspection)



So now to replace more caps and mod it.
I have one more defective K8N-DL so I will test best modding temps there first.

 

[digitalbath]

Missing capacitor CE75? Nice work.

 

[Sparksnl]

Yes, sharp as always.

 

[stunned_guy]

Funktionieren alle RAM-Steckplätze noch ? Auf dem 2. Bild sehen die mit Wärmeleitpaste beschmiert aus und teilweise zerstört. Kann aber auch täuschen.

 

[Sparksnl]

That is also sharp, I removed it before testing it. It was not enough to do any damage.

 

[Sparksnl]

So I have been busy today... Both CPU's have no Coldbug post issues.
Some results for LN2.

FX-60 CCB2E 0536MPMW
Best run at -45C and scaling better then expected. I expected it to do a 200MHz lower for validation (it was not that good on air). Easy bench CPU, no issues at all.
3722MHz validated

FX-60 LCBIE 0611SPAW
Best run at -20C to -26C. Running Ok but sadly it can not go cold enough to do more damage.
3733MHz validated and HWBot.

It can do more, but validation is not possible: 3746MHz max


PI 1M quick bench: 24.234s


Edit: this just came in

 

[stunned_guy]

0543 FPMW: eins der ersten richtig guten Steppings Viel Erfolg und gut Glück !

 

[Sparksnl]

Yes I expected it. I noticed more then one good 0543 cpu and since it is BC it could have a good coldbug. Probably an FX sold as 3700+.
It came from an Acer pc I noticed, it still has it’s label. So probably not overclocked as well.

 

[Sparksnl]

Today a Tyan K8WE S2895 ((S2895A2NRF) came in.
After (cleaning it and) replacing 7 capacitors (Among them the 16v's, they were bad. That was a first) it workes fine.
It cost me €15, so this is above expectations.

It came with 2x Opteron 270 and 8x Kingston 1GB (Samsung ZCCC).

 

[Tzk]

1GB Samsung ZCCC

I also got these, curious how far you can push them. I run them on my dual 462 build, at a low-ish 150Mhz 2-2-2-5

 

[Sparksnl]

I will probably test them at some point in my Asus board. But 2-2-2-5 sounds great.

 

[Tzk]

Don't expect much. ZCCC isn't great and thus won't run 2-2-2 anywhere near 200MHz. It works on my machine because it's underclocked.

 

[Sparksnl]

I kind of expected it (to not be that good).

Edit: I noticed my Tyan board can disable a core. Would it be possible to add this feature to other motherboards?

 

[Raiden Zero]

FX-60 CCB2E 0536MPMW
Best run at -45C and scaling better then expected. I expected it to do a 200MHz lower for validation (it was not that good on air). Easy bench CPU, no issues at all.

Looks pretty similar to my 0536RPMW. Also easy to bench on single stage and does 3.4GHz Cinebench, so probably around 3.5GHz for 1M

 

[Sparksnl]

Looks like the 2005 batch is a bit better then the 2006 batch (for B2).

Edit: 1 more pic of the Tyan board.

 

[bschicht86]

Looks pretty similar to my 0536RPMW.

Taugen die unter Luft was? Weil ich habe ebenfalls einen FX60 0536RPMW. Und einen 4800+ 0536XPMW.

 

[Sparksnl]

Results so far for my new 4000+ (KAB1E 0626FPMW), boot tests:

3096MHz 1,3000v
3204MHz 1,3750v
3301MHz 1,4750v

 

[McZonk]

Results so far for my new 4000+ (KAB1E 0626FPMW),

Quicktested mine for Max FSB as follows:

nf4 Expert, AMD Wraith Prism, BIOS 1,50V (1,472V real), Multi 10, LDT 3, RAM5/6, Prime blend 5 Min auf 5 MHz Bclk

Results in 305 MHz and is my second best chip on air.

 

[Sparksnl]

Max fsb is best tested at a low cpu multi and fsb/ldt at 2.0 or 2.5 (could make a difference one or the other.

I seem to have enough good 4000+ cpu’s.

Edit: my nr.4 San Diego 4000+ needs 1.59v to boot 3.2GHz

 

[McZonk]

My wording was a little imprecise here.

I test for max. FSB in conjunction with Multi 10, so no test for absolute max. FSB but rather for the Max Frequency at a certain voltage.

 

[Sparksnl]

So today was another LN2 session. Next on my list was my Athlon 64 FX-55 Clawhammer (AAA2C 0437UPBW).

I only have 1 picture.


With 1.64v for CPU and a multi of 5x I had a max FSB of:

Pot: -50C / Sensor: -45C
FSB/PLL: 2.5x - 380MHz FSB
FSB/PLL: 2.0x - 370MHz FSB

Pot: -80C / Sensor: -72C
FSB/PLL: 3.0x - 365MHz FSB
FSB/PLL: 2.5x - 365MHz FSB
FSB/PLL: 2.0x - 335MHz FSB

Pot: -140C / Sensor: -125C
FSB/PLL: 3.0x - 305MHz FSB
FSB/PLL: 2.5x - 315MHz FSB
FSB/PLL: 2.0x - 300MHz FSB

So Clawhammer does not seem to have any max FSB issues (as that max FSB is too low to reach a high CPU MHz).
The only thing I saw was that just like San Diego and Toledo it has a best temperature vs max CPU MHz scaling issue.
Best run temp was -53C to -63C (great temp for my SS, but still to low). Where it had a max validation of 3596MHz.
It did not have any coldbug post issue up to -144C (I did not test any colder).

Next year 2 more FX-55 Clawhammers on cold.

Edit: At -140C the max frequency to be able to boot at 1.64v was 2800MHz. So to much cold was given.

 

[bschicht86]

Ich habe von @Ben Psyko einen Haufen CPU's bekommen, wo mindestens ein Pin fehlte. Nachdem die wieder reapriert waren, habe ich mich mal der schönen Steppings ans austesten gemacht und habe meine mit dazu genommen. Auf einem A8V Deluxe mit einem billigen Alu-Kühler ohne WLP kam folgendes heraus: (Alles 4600+)






Erklärungen:
Mit 1,5V hab ich mich an die 3GHz ran getastet, danach die Spannung runter gestellt.

"3,0" = Taktfrequenz
nichts dahinter = hat memtest bei diesem Takt eine ganze Runde fehlerfrei überlebt
"E" = Fehler in memtest
"X" = Hat sich bei memtest und davor aufgehangen
"12x250" = Multi und FSB
"1,5V" = Spannung lt. BIOS

Ergebnisse:
LCBIE 0628CPMW 3,0 1,35V
LCBIE 0627TPMW 3,0E/X 1,35V
LCBIE 0627TPMW 3,0E 1,5V
CDB3E 0607XPDW 3,0E 1,5V
LCBIE 0627TPMW 3,0X 1,5V
CDB3E 0607XPMW 3,0X 1,5V
CDB3E 0607XPMW 2,94 1,5V
CCBIE 0608BPMW 2,94X 1,5V
LDBBE 0601TPMW 2,82 1,5V
CDBHE 0531SPMW 2,76X 1,5V
LDBFE 0613RPMW 2,76X 1,5V
LCBIE 0627TPMW 3,0 1,45V
CDB3E 0607XPMW 2,82 1,5V
LCBIE 0628CPMW 3,0E/X 1,5V
CCBIE 0608APMW 2,94E 1,5V
CCBHE 0531BSPMW 2,88X 1,5V

 

[mrcoconut]

Anhang anzeigen 954326
Anhang anzeigen 954327
Anhang anzeigen 954328
Anhang anzeigen 954329

Noch so ein Graffitikünstler 🙃

VG 🙂✌️

 

[bschicht86]

Noch so ein Graffitikünstler 🙃

Ich war nur die dünne Schrift. Und die lässt sich mit Bremsenreiniger wieder fort wischen.

 

[mrcoconut]

Ich war nur die dünne Schrift. Und die lässt sich mit Bremsenreiniger wieder fort wischen.

Oder Isoprop 🤪

VG 🙂✌️

 

[stunned_guy]

Da sind ja Toledos und Manchester gut gemischt. Sind die CPUs nur zum Reparieren zu dir gekommen oder sind die jetzt deine ? Der Markt an X2 4600+ ist irgendwie wie leergefegt.

 

[bschicht86]

Zum reparieren.