DFI 925XE-T2V @ 7.2Ghz

[Del Gro Badeu]

have cpu-z open before running super pi and monitor while running to see if it comes up while super pi is running?

does wcpuid realtime clock checker work?

normally, no ... according to the_mad, only CPU Z 1.27 is detecting this multiplier changing

[xsecret2]

We used various tools for testing that the speed was correct and it came out correct every time. If I select CPU #2 in 1.27 it shows the correct speed.

Franck just uploaded a new version of CPU-Z 1.27 that solve this issue. This bug is really easy to understand :

When starting, your CPU is in IDLE mode (C1E => 14x), so CPU-Z compute frequency for CPU #1. Then CPU-Z compute frequency for CPU #2. At this time, your CPU is *NOT* in idle C1E mode because of the computation of CPU #1 he just did. So, it reports an higher, non-idle frequency. I have a F41 Prescot and I run some tests. You can find these attached and you will see about CPU-Z V1.26 / Crystal CPUID and other tool that does not report the correct speed in idle with PRescott F41.

By the way, all you need to do is to disable C1E (if this doesn't work on the DFI, that's a BIOS Bug because it works on Asus P5AD2-E), or add a little CPU load when you output the CPU-Z validation file, and all the world will be able to saw your awesome overclocking

[death metal]

xsecret2, thanks for the clarification...and i hope we see you dropping by here more often =).....if i remember correctly, it's at x86 where i've seen very early OC findings of Pentium M =)....

[xsecret2]

Haha yes We use to leak some not-yet-announced CPUs like P4 630 some weeks ago (Smithfield coming soon) and we helped DFI a bit for the 855GME-MGF

[death metal]

hehehe, yeah, DFI's mobo is an instant hit =)....i'd love to see those Smithfield reviews...and if only better translators would be available, your articles will even be more entertaining...i can't believe the word "caramel" will always be in your article when translated ....i know it's not supposed to be caramel, but then again, either i try so hard to read the translated page or completely skip reading it since i don't understand a single word of it hehehe....

[Exerqtor]

Yo everyone!! Im new here!

Congrats fugger

[Charles Wirth]

Let me get this right, opening CPUZ is not enough of a load to bring CPU #1 to correct speed but clicking on tab #2 is?

Firing up and running test for you then going on web taking screen shot of CPU incorrect then uploading to your site was not enough load to get out of C1E state but clicking on cpu #2 is?

Every time I opened 1.27 before and after tests it was always incorrect.

Methinks a) C1E is not working on my board or b) another bios rev needs to be done as my bios is from original 925X.

FMB15 enable works correctly but that is different and made to force retail CPU to lowest multiple for 266FSB default operation.

I will have it open for super pi and get a screen shot.

So far the best I can prove is 1.27 was reporting the wrong speed. You guys are top notch and I am sure we can figure this out.

[Supertim0r]

me <3 intel

[STEvil]

Fugger - I think what he meant to say is CPU-Z calculates the frequency of the CPU quick enough that the CPU isnt up to speed until it starts calculating the "second" CPU's frequency.

[Cerberus2k7]

haha, I saw this in the news box over at OCF and I said to myself, "Only someone from XtremeSystems could do that." and sure enough; the link too me here. kick ass work man!

[xsecret2]

Let me get this right, opening CPUZ is not enough of a load to bring CPU #1 to correct speed but clicking on tab #2 is?

That is correct : Opening the first version of CPUZ 1.27 is not enough of a load to bring CPU #1 to correct speed but clicking on tab #2 is, because of the computation just done of CPU #1 speed.

BTW, please download again CPU-Z 1.27 here : http://www.cpuid.org/download/cpu-z-127.zip

This new 1.27 version solve this bug and both logical CPU will report the same, correct speed.

[i found nemo]

you crazy mother fugger, lol had to say that, good job man wish i could hit half that lol

[Charles Wirth]

Seems to be working as intended now.

rebooting up upload monster result.

[Charles Wirth]

Result uploaded. bring on the clockgen

http://valid.x86-secret.com/show_oc?id=2892

[GRAFiZ]

Result uploaded. bring on the clockgen

http://valid.x86-secret.com/show_oc?id=2892

Whoa! I like what I see... keep it up bro!!!

[[XC] moddolicous]

Good job so far, but u have to get it up to at least 6.5ghz in windows to hold the WR!

[dippyskoodlez]

Good job so far, but u have to get it up to at least 6.5ghz in windows to hold the WR!

I think he knows that.. he set the record didnt he?

[shadowing]

Come on. Beat the WR!!

[Charles Wirth]

Franck says I should be able to test tomorrow. I had horrid contact when popped up that 5.8Ghz. I didnt care about max for that one, just wanted a high speed SS to post.

I wish it shown the voltage. I was at 1.53v @ 5.8Ghz

[Jrocket]

I wish it shown the voltage. I was at 1.53v @ 5.8Ghz

That is incredible! How high have you ever run that chip?

[Charles Wirth]

in windows pi @ 5.9Ghz 328FSB 1:1 3-2-2-8

[Jrocket]

Actually... I meant what is the highest vcore you have run that chip? However that is very impressive, especially being able to run the ram that tight at that speed.

[Sup3rman]

Amazing...Amazing...Amazing...

[imbecile]

Whoa, awesome. great job. what kind of cooling system do you use?

[zA.Gosu]

^
I
I
I

NEWBIE!

LOL. Sry. He uses a coupla cascades. It is SUPER COLD, anywhere below -50C