Discussion in 'News' started by btarunr, May 3, 2013.
8Ghz is better
Invalid link = fake.
But you cannot. Too bad.
Yeah that 2.56V is fake, cpu-z bug.
One user from OC kingpin cooling who posted this info at guru3d forums said so.. I take his word for granted, becasuse he's no noob ^^
And if you look at that quad 6.2Ghz @ only 1.21v
Anyway I can't wait to get my hands on 4770K, looks like 5.4ghz won't be a problem, probably 6.2 ghz for sure, woot! :]
Am i stupid, or this was single channel only? LoL
Not stupid... but it isnt relavent either.
Sure it is, easier on system... Dual channel or GTFO! LoL
Pfft. More dimms and moar power! Quad-channel or get out.
I'm on triple channel so i'm cooler. Ménage à trois.
Right, but that (clearly) isnt a 24/7 thing... so, who cares how many DIMMs it ran. Just like who cares how many cores it was running to get there (as people moan "zOMG thats only on two cores....").
Things are a bit different with Haswell, it seems, compared to SB and IB overclocking.
I think I should also point out that last I heard Haswell's max multi was 63, not 77 on k-edition CPUs.
I wish I caught that sooner even though I already knew the vcore was bogus, now I know the entire thing is bogus.
"With unlocked "-K" chips, you get the freedom to step up base clock multiplier for the CPU cores all the way up to 80.0x for 100 MHz, up to 64.0x for 125 MHz, and up to 48.0x for 166 MHz; which if used right, could result in some awesome CPU clock speeds in the neighborhood of 8.00 GHz."
Did anyone take note that this was done with just hyperthreading turned off (and a single DIMM, but still)? Most of these extreme overclocks only allow 1 or 2 cores to reach speeds in excess of 6-6.5GHZ, mostly due to stability. But this little gem did it with all 4 physical cores!
Maybe that's why CPU-z reads 2.62v, hmmmm.
So I gut into the multiplier deal a bit more and I don't see anything that confirms those multiplier numbers. Those are from the rumor mill according to TomsHardware.com and hasn't been confirmed. I'll save the speculation until we get some real facts.
Or maybe it means that it is fake...
Of course it is a cherry picked sample, that is pretty given.
And have you ever seen some of the extreme shit these extreme overclockers do to their motherboards?
They add phases, bridge phases from other motherboard into one, etc.
Have you never seen that?
This was a suicide run, obviously, and they'll do crazy shit just to get a CPUz screenshot before the system dies.
Yeah, that's how they double current, not voltage.
The point is that you would need to at least replace all of the caps with ones that support twice the voltage (which could easily translate to caps that are twice in size.) Now that you've swapped out the caps you would need to hack the BIOS because the caps will burst if you even attempt to run VRMs at this voltage. Also if the capacitance of any of the different phases changes, you need more than a custom bios, you need to update it to handle the hardware changes you've made.
Also the addon board you speak of I've only seen for Kepler-based GPUs since voltage control is locked. I've yet to see this for the motherboard and CPU power phases, but once again there is a different reason for that.
It would have died the second it booted with a voltage like that. LN2 doesn't let you overcome physical limitations like the breakdown voltage of the caps. You'd pop the caps (haha!) before the VRMs could even charge.
Interesting as I said in the AMD thread not a single nay-sayer about an Intel chip clocking high in a completely useless LN2 infused scenario.
Maybe there wasn't but my opinion is, in a nutshell the same. If they cripple the CPU just to get a huigh clock it' worthless. So disabling L2/L3 caches, HT and cores falls into that category.
CPU-Z says 4 cores, just with that it's already a lot more impressive than that FX record OC that turned off 7 of the 8 cores.
It looks like Intel is trying to sway public opine because Haswell only delivers 5%-10% improvement over Ivy Bridge, which was barely better in performance than Sandy Bridge.
But if it can be overclocked to 6GHz so easily, with an even higher IPC than Ivy Bridge, it could cause quite a stir in the enthusiast world. Think of all the people who will be rushing to get one of these, even people already sitting on the top end. There's a lot of people like me who are sitting on much older yet still very capable hardware who may want to upgrade as well.
Yeah, pair that with overclocks in the 5Ghz-5.5Ghz range and you have a substantial increase in performance. I suspect Haswell will overclock better than IVB does.
Well done. Excellent research.
From TechPowerUp three weeks ago:
I said last time I heard and said nothing about doing research. I stand corrected but there are nicer ways to come about saying it. Don't be a smart ass.
So true. Does this example work better? Say change...
i.e. turning the comment into a question rather than accusing rtiueuiurei of falsifying their benchmarking results based on your incomplete knowledge base.
I don't know, but if you're throwing out words like "bogus" and "now I know", wouldn't you at least spend 30 seconds making sure you were standing on solid ground ?
I already said that I got it wrong and you don't have to go about saying how else I could have said it. All you're doing is fishing for me to lose my cool and it's about the most immature thing someone can do on these forums can do. How about you stop trying to piss me off already. If you can't act like a civil human being you can take you bantering and self-righteous attitude to GN where it will be welcomed.
Separate names with a comma.