• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

4670k - bad core?

B

bipolar

[H]ard|Gawd
Joined
Jan 6, 2003
Messages
1,184
I think one of the cores on my 4670k might be bad. I don't know if that's even possible -- it seems like a cpu would either work or not -- but this is why I think it might be the case.

So, I bought a Haswell + new motherboard and installed it, everything seemed to work fine. Before overclocking I decided to run a basic benchmark just to see any improvements over my old q6600. I used HyperPi, which is a version of SuperPi that runs concurrent tests on all cores. So I benched pi to 1 million places. Cores 1, 2, and 3 were all within .1 second of each other. Core 0 was twice that of the others. I thought it might be because Windows was running on that core and so disabled as much as I could, but the results were about the same.

Well, maybe HyperPi is buggy or something, no big deal. Overclocked (only to 4.1, but I haven't messed with voltage yet and my hsf is nothing great) and reran HyperPi, with similar results (lower overall times, but core 0 was still twice that of the others).
Then I ran 3DMark and compared my numbers with similar systems. My graphics scores were fine, but the physics score was around 6300 and every other 4670 that was even close to my cpu's speed (even if they were several hundred Mhz slower) had a physics score of 9000+

It's not throttling -- Core 0 is one of the two coolest running cores in fact. Core 1 has peaked at 72 degrees under 100% load for long periods, but Core 0 has never broken 66.

There's an Intel utility that tests the cpu and it didn't give any errors, but I don't know what it's testing for or if it would detect a cpu that works but just more slowly.

I mean, it's definitely faster than my 3.2 ghz q6600, but I can't help but be disappointed that one core seems so slow and my numbers aren't what they could / should be.

Any ideas?
 
Try running Cinebench?

Look at the task manager... are all the cores being used 100%?
 
Ran it and got 524, with 99% cpu usage in Task Manager. I don't know what that proves per se, as I guess a broken core could still be working as hard as it can even if it's 'damaged' or w/e
 
SolidBladez said:
Is this on a fresh OS? Have you updated the motherboard BIOS?
Click to expand...

Fresh OS, yes. I've tried to update the BIOS and have run into problems -- the Windows utility can either point to a local file or get the updated BIOS from the internet. I've tried both ways and it looks like it programs it successfully either way (and the version it is allegedly programming is definitely newer than the one that is on there), but the BIOS version does not update (per both the ver info in the BIOS and per CPU-Z). I have a support ticket in but haven't heard back yet.
The BIOS can also be updated via a usb thumbdrive but I don't have one so the Windows utility was the only means available to me.

It's a Biostar z87x 3D / Hi-Fi, for what it's worth.
 
You must log in or register to reply here.
Back
Top