Intel's Core i7-7700K @ 6.7GHz On LN2

H

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
If this Facebook post is to be believed, one of Intel's upcoming Core i7-7700K processors was able to hit 6.7GHz on liquid nitrogen. Given that these chips are reportedly running at 4.2 GHz base and 4.5 GHz boost, that's a pretty decent OC, LN2 or not.
 
I don't understand why people do this.

Bigger epeen?

I mean, a chip cooled with LN2 isn't going to be practically usable...
 
Zarathustra[H] said:
I don't understand why people do this.

Bigger epeen?

I mean, a chip cooled with LN2 isn't going to be practically usable...
Click to expand...
LMAO Zarath! As long as you been here and as many posts you've made, you dont know why people shock their systems w/ LN2? Oh, you answered your question. Epeen and to show what the chip/tech future can be capable of.
 
munkle said:
Why do people by powerful cars? We all have to go the same speed limit.
Click to expand...
They're still practically usable for racing. An LN2 cooled OC record is as usable as a car that has 1 wheel and explodes if you try to rev the engine. It's like building a car, for the dyno bench only, that is never meant to be driven.
 
Steve said:
If this Facebook post is to be believed, one of Intel's upcoming Core i7-7700K processors was able to hit 6.7GHz on liquid nitrogen. Given that these chips are reportedly running at 4.2 GHz base and 4.5 GHz boost, that's a pretty decent OC, LN2 or not.
Click to expand...
The definition of a decent OC has changed rapidly over the past years. 60% would be a decent (not great) OC with air if we go by historic figures. And we don't even see if it has all cores running or only one.
 
M76 said:
They're still practically usable for racing. An LN2 cooled OC record is as usable as a car that has 1 wheel and explodes if you try to rev the engine. It's like building a car, for the dyno bench only, that is never meant to be driven.
Click to expand...
I'd buy that car
 
M76 said:
They're still practically usable for racing. An LN2 cooled OC record is as usable as a car that has 1 wheel and explodes if you try to rev the engine. It's like building a car, for the dyno bench only, that is never meant to be driven.
Click to expand...

So a drag racer...
 
And I don't think they stability test LN2 Overclocks. To me, if they aren't stable, they are useless. A bit like a dragster.
 
Colour me ignorant here, but that Clock is hit on LN2 by using only 1 core or using the entire chip?

To me, if it is the former, it's a completely fruitless exercise for practical purposes beyond being a hobby doing it "just because they can".
 
i7Baby said:
And I don't think they stability test LN2 Overclocks. To me, if they aren't stable, they are useless. A bit like a dragster.
Click to expand...

None of those high LN2/DICE crazy overclocks are stable under any sort of benchmark. You practically go as high as you can with the clocks, go into windows and if you're lucky enough your PC won't reboot or crash while submitting a CPU-Z validation online then you have that record. It's been this case since the dawn of overclocking.

The point of overclocking is to demonstrate how far the boundaries of the CPU/processor itself can be pushed and other reason is for entertainment/fun/to experience it etc etc.

Same concept of idea with IBM demonstrating a 100 core CPU almost 10 years ago. Does it work? No, but it's just to demonstrate the engineering prowess of the company has achieved or the extreme end capability of a piece of tech.
 
Things like this use to be more fun when everyone would chip in on a dewar, pass out some beer with the thermoses, and it was more of a party/competition. Just seeing someone doing it in a video for pure solo epeen just does nothing for me.

I also have a pretty restrictive definition of stable that I'm sure this wouldn't meet. It's like throwing twin turbos on a stock, high compression, engine and turning up the boost. That first burnout is cool until the engine goes tits up.
 
The trend isn't looking good.. a 59% overclock using LN2 isn't "pretty decent" when compared to some other LN2 overclocks previously mentioned in [H]..

2013 - Asrock managed to overclock a 3.5GHz Core i7-4770K to 6.5GHz = 86% increase

2012 - Core i7-3770k running at 7.06GHz = 102% increase

2011 - Kyle at AMD Austin watches a FX 8150 on LN2 overclocked to 8GHz = 158% increase

B0_EHXXq_CYAAo_Vyd.jpg
 
I was always part of the crowd that was more impressed with 24/7 systems.

I had a watercooled peltier system for a bit that ran at -5C. I dismantled it though when I got laid off. I didn't want to wreck $4k in parts (Titan X SLI + 5960x). It cost a fortune to run too. 4kW that I actually dumped the heat into my pool. Wife was like, "wow, how is the pool 80*F in April!?"

Some others use multiple phase compressors. Those are loud and you can't point the heat where you want it... I guess I could water cool it instead of air cooling.
 
munkle said:
So a drag racer...
Click to expand...
A drag race is still racing. Usually many rounds in one competition. An LN2 PC running on 1 out of 4 cores has no purpose whatsoever. You couldn't drag race a car that runs on 1 out of 4 cylinders.
 
To me, the purpose of doing this is two fold:

1. Because it's a hobby, and people with the disposable income to do so, should be able to.
2. To provide valuable feedback to the CPU manufacturer that can be applied to future designs for achieving higher factory clock speeds while increasing IPC, decreasing cooling requirements, and decrasing node size with each new processor generation.
 
DejaWiz said:
To me, the purpose of doing this is two fold:

1. Because it's a hobby, and people with the disposable income to do so, should be able to.
2. To provide valuable feedback to the CPU manufacturer that can be applied to future designs for achieving higher factory clock speeds while increasing IPC, decreasing cooling requirements, and decrasing node size with each new processor generation.
Click to expand...
Exactly. Just because I would never do this doesn't mean that other people with the means and connections shouldn't bother. Now I'd be a lot more impressed if it was over 8 GHz, but whatever.
 
M76 said:
A drag race is still racing. Usually many rounds in one competition. An LN2 PC running on 1 out of 4 cores has no purpose whatsoever. You couldn't drag race a car that runs on 1 out of 4 cylinders.
Click to expand...
I like your point, though I did once drive my old 454 pickup around the block on only 4 out of 8 cylinders after retorqueing the damn header bolts and forgetting to put the spark plug wires back on. It even ran pretty well considering.
 
M76 said:
A drag race is still racing. Usually many rounds in one competition. An LN2 PC running on 1 out of 4 cores has no purpose whatsoever. You couldn't drag race a car that runs on 1 out of 4 cylinders.
Click to expand...

The purpose of these world record overclocks is to try and push the speeds as high as possible, no matter the amount of cores or threads it has. This also includes disabling cores or other functions or components to try and reach said speeds.
 
Matthew Kane said:
The purpose of these world record overclocks is to try and push the speeds as high as possible, no matter the amount of cores or threads it has. This also includes disabling cores or other functions or components to try and reach said speeds.
Click to expand...
I know the goal. I question the real world value. I personally don't care about any overclocking result that is not at least as stable to run benchmarks. It's a meaningless number, to stay at the car analogy it's like trying to find out how much rpm the crankshaft can take without pistons.
 
M76 said:
I know the goal. I question the real world value. I personally don't care about any overclocking result that is not at least as stable to run benchmarks. It's a meaningless number, to stay at the car analogy it's like trying to find out how much rpm the crankshaft can take without pistons.
Click to expand...
There is almost zero real world value for individuals other than bragging rights and learning some of the intricacies on how to replicate and evolve the process for current and new chips. However, the manufacturers might glean some info off of it. The manufacturer may or may not have done similar testing already, but the overclocker's tinkering gives them more data to work with on what their product can do under certain situations.

Think of it like car racing. The manufacturer looks at the racers and thinks, "these guys took our 300hp V8 and made it into a 1200hp V8. Granted, the thing blew up after a quarter mile but I think we can improve on what these guys did since it's our engine. I bet we can not only get 600hp out of it, but improve the fuel economy too." The racers then take the new 600hp engine and get 1600hp out of it... rinse repeat.
 
You must log in or register to reply here.
Back
Top