Haswell Voltage

[ProKoN]

You posted it on this forum. 

 

https://docs.google.com/spreadsheet/ccc?key=0AlC81MjwelBgdEZNV3l6aHl1eUNwSUR4Rml0MXMzN1E&usp=sharing#gid=0

 

You may be confused. The results posted there are merely benchmarking fun. this thread has nothing to do with stability.  please post your best results. no one cares about temps or voltages unless you  are throttling lol. this is a great thread and is well maintained by @Jumper118.

 

no where in my guide do i reference those results.

 

My guide promotes stability. and is based off of a modest overclock. 4.4 ghz with a starting voltage of 1.25v and always promoting voltage reduction.

 

im not a believer in aida64 since it is a paid for product and it can not cope with adaptive voltage, id rather use prime95.

 

and just lol in general.

[FireFox]

More evidence that says Sandy Bridge is superior

[Jumper118]

More evidence that says Sandy Bridge is superior

Yep.  

There are some serious walls of rage text here. I would just like to say that cinebench takes an entire minute or less to do and that most high scoring people wouldn't and don't run there cinebench clocks 24/7 because there is a balance between performance, heat output and power usage, the higher clocks normally require a lot more voltage for maybe an extra 100mhz which isn't necessary in normal use.

[deathjester]

You may be confused. The results posted there are merely benchmarking fun. this thread has nothing to do with stability.  please post your best results. no one cares about temps or voltages unless you  are throttling lol. this is a great thread and is well maintained by @Jumper118.

 

no where in my guide do i reference those results.

 

My guide promotes stability. and is based off of a modest overclock. 4.4 ghz with a starting voltage of 1.25v and always promoting voltage reduction.

 

im not a believer in aida64 since it is a paid for product and it can not cope with adaptive voltage, id rather use prime95.

 

and just lol in general.

 

Prime 95 which doesn't even test the new instructions. Long stability tests were pretty awesome on my e8400 that ran cool. Were also great on Sandy Bridge. Haswell? Different story. Might shock you but Prime 95 (even blend) goes bonkers after round 1 on air cooling. An even bigger shocker? Most people don't have dual rad cooling like you, and even when they do? If they are using prime 95 at 1.6v (lol), I want to see video of it with temps.

 

I hope they are running the chip naked without a lid or something really elaborate. The chip will still die, no matter what the temps, but I still want to see it. Maybe you should try it on yours. Make a video of it. You can send another chip back to Intel (if you already didn't use up the 1 chip thing when you broke the first one).

 

Anyways I am done conversing with you. It has been fun. I appreciate you in instructing me how to stability test duocore. I knew it back then, but I didn't know it was still relevant. Thanks for the tips. All the motherboard companies and tech sites recommending AIDA 64 are obviously bonkers, and Asus making realbench with their own money to help everyone was obviously wasting their time.

 

P.S. One round of prime blend is still fantastic for testing a memory overclock quickly or good to throw in with other stability tests as a SHORT test. The other tests (non blend) are pointless on Haswell and you are running stupidly high temps for no reason.  All you are doing is seeing if your chip runs at temps you will never see, with old instruction sets. Your chip should never see those temps anyways which makes the whole test silly.

 

If you want to spend days, doing what should take hours finding stability? Go for it. It's your time man. It isn't the best way, and isn't recommended by anyone, but if you are comfortable with the old way of testing and you have dual rad water? More power to you.Feel free to "lol" more at me. I have provided multiple proof from tech sites, MB manufacturers and we already know what Intel says. You against the world. Have fun in your quest.

[ProKoN]

here is version of prime95 the stresses fma3 and avx2.

 

http://www.sendspace.com/file/mtfcor

 

if you look at any of my previous posts. i do Not recommended using prime95 for haswell,

 

 

 

 

anyways my voltage recommendation was already posted by @mgsstar

[flibberdipper]

Take the clock down so that it's stable at 1.35v. Then dance in the rain of profit.

[deathjester]

here is version of prime95 the stresses fma3 and avx2.

 

http://www.sendspace.com/file/mtfcor

 

if you look at any of my previous posts. i do Not recommended using prime95 for haswell,

 

 

 

 

anyways my voltage recommendation was already posted by @mgsstar

 

Cool man. Run that Prime with new instructions for 8 hours on a air cooler like a evo 212 that most people have and see what happens even at low voltage like 1.25v.

 

Better yet. Run it past 1.4 V on a H100i which is the "premium" option the average person springs for. Make a video and show temps.

 

I will eagerly be awaiting that. You have a water cooler that cost 140-150 bucks, and act as if everyone should test like you do. To put it simply? They cant. The temps aren't safe for your suggested "at least 8 hours of testing" even past 1.3 on most of these all in one water kits, let alone 1.4 and your "anything under 1.6v" is fine. You seem to think all haswell batches run at the same temp. They don't. Some 4770k's hit damn near throttling on a stock cooler in a render. 

 

Is this with a delid as well? Most people aren't doing that either.

 

Your advice is suited to people with delidded chips on a 150 dollar cooler or a custom built loop and you have no evidence that 1.4v is safe at all (even with low temps) and people have had a chip die in a month past 1.4v with low temps on custom loops.

 

Your voltage advice is = to Sandy overclocking, which is quite laughable to be honest. The only way anything close to 1.6v would be ok on temps is if you considered farmville gaming and MS paint load work. 

[flibberdipper]

Cool man. Run that Prime with new instructions for 8 hours on a air cooler like a evo 212 that most people have and see what happens even at low voltage like 1.25v.

 

Better yet. Run it past 1.4 V on a H100i which is the "premium" option the average person springs for. Make a video and show temps.

 

I will eagerly be awaiting that. You have a water cooler that cost 140-150 bucks, and act as if everyone should test like you do. To put it simply? They cant. The temps aren't safe for your suggested "at least 8 hours of testing" even past 1.3 on most of these all in one water kits, let alone 1.4 and your "anything under 1.6v" is fine. You seem to think all haswell batches run at the same temp. They don't. Some 4770k's hit damn near throttling on a stock cooler in a render. 

 

Is this with a delid as well? Most people aren't doing that either.

 

Your advice is suited to people with delidded chips on a 150 dollar cooler or a custom built loop and you have no evidence that 1.4v is safe at all (even with low temps) and people have had a chip die in a month past 1.4v with low temps on custom loops.

 

Your voltage advice is = to Sandy overclocking, which is quite laughable to be honest. The only way anything close to 1.6v would be ok on temps is if you considered farmville gaming and MS paint load work. 

You should NEVER, EVER, EVER run Prime95 on a Haswell chip.