• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

My g3258 overclock

aznb0b

Member
http://imgur.com/tH6dv4Q

I believe I can bump the multiplier up some more since I haven't even gone over 61c during stress tests.

This is on my msi z87-g41 with a corsair h60 cpu cooler.
This was also done using OC genie and msi command center. IBT STILL shows me in the log at a lower clock speed but EVERY other program is showing me at 4.6ghz so Im satisfied with the majority!
 
One of my G3258 runs at 4.5GHz on stock cooling. The other runs at 4.4. I dont think I'd be satisfied with that OC on a cooler of that caliber.
 
Your temps are good, though the throttling during IBT is problematic. I think I asked in another thread (don't recall if I got an answer), but what is your throughput in IBT? That can tell us quite a bit about what clockspeed you're really running in that test.
 
IBT runs an unadultered copy of Intel's linpack benchmark. It does some sort of mathematical calculations that require an enormous amount of system memory (you have the option to use all available memory if you like). Every time it processes a block or whatever you call it, it reports throughput in the form of gigaflops per second, or Gflops/s. For example, my tri-core chip tops out at maybe 40 Gflops/s if I push it hard. An i5 at around 4.4 should hit maybe 90 Gflops, possibly more. If you are getting somewhere between 45-50 Gflops/s then you are probably not throttling during IBT. Anything less than that and you should start questioning your overclock.
 
IBT runs an unadultered copy of Intel's linpack benchmark. It does some sort of mathematical calculations that require an enormous amount of system memory (you have the option to use all available memory if you like). Every time it processes a block or whatever you call it, it reports throughput in the form of gigaflops per second, or Gflops/s. For example, my tri-core chip tops out at maybe 40 Gflops/s if I push it hard. An i5 at around 4.4 should hit maybe 90 Gflops, possibly more. If you are getting somewhere between 45-50 Gflops/s then you are probably not throttling during IBT. Anything less than that and you should start questioning your overclock.

On my Ivy i5, it's close to 120Gflops at 4.6. It seems to scale pretty linearly with clock, and I'd expect Haswell to be a bit higher per clock.
 
If you left the power limiter in the bios at stock the cpu might clock back when you run something like intel burnin test.
 
On my Ivy i5, it's close to 120Gflops at 4.6. It seems to scale pretty linearly with clock, and I'd expect Haswell to be a bit higher per clock.

Hmm, then yes it should do better than 45-50. Not sure where I read about a Haswell i5 only turning in 90 Gflops/s, but that seems to be in error.

So maybe his G3258 should be closer to 60-70 Gflops/s.
 
Last edited:
Okay, AVX makes a difference.

Well aznbob, now you know what throughput is and you have one reference point showing you what you "should" have, though 34.13 Gflops/s sounds kinda slow actually. I beat that with a K10.5 tri-core. Maybe I need to be using a newer version of linpack?
 
Back
Top