9900KS...

[Besty]

Thought the next one wasn't until 18 months away?

https://forums.overclockers.co.uk/posts/33055446

For the Extreme Series 10th Gen, we will likely see reviews before this Christmas. Launch is very close.

 

[TangoEchoAlpha]

9900ks must be getting close - 9900k has had a further £20 discount on OCUK.

 

[Woodsta888]

It will be interesting to see the cooling required for these chips. Shouldn't be too bad as they are binned not only to reach 5Ghz but also under a voltage threshold so less heat produced. Will top air coolers be enough or will a 360 AIO be required?

 

[Grim5]

From what I've read these chips are binned to run 5ghz all core at 1.3v OR lower.

I think a NHD15 can handle that

 

[Besty]

It will be interesting to see the cooling required for these chips. Shouldn't be too bad as they are binned not only to reach 5Ghz but also under a voltage threshold so less heat produced. Will top air coolers be enough or will a 360 AIO be required?

personally, i think a 360 is enough for gaming however, the fan speeds on them will be a problem unless you have good sound dampening.

 

[Deleted User 456458]

It will be interesting to see the cooling required for these chips. Shouldn't be too bad as they are binned not only to reach 5Ghz but also under a voltage threshold so less heat produced. Will top air coolers be enough or will a 360 AIO be required?

Comes down to how many amps and load voltage you’re pulling. Your cooler can either dissipate x amount of heat or it can’t.

Amps * load voltage = watts

watts = heat

150amps is not unreasonable for an oc’d 9900k. For that you need a 360 aio at full fans for sustained workloads.

150 * 1.2v = 180w
150 * 1.3v = 195w

The above matters a ton.

200amps needs a direct die with a custom loop.

 

[Cooper]

personally, i think a 360 is enough for gaming

360 rad for gaming?

My dark rock 4 can keep a 9900K @5Ghz @1.35vlts under 60c in gaming. For just gaming, you don't need any kind of water cooling for a 9900K.

 

[Woodsta888]

360 rad for gaming?

My dark rock 4 can keep a 9900K @5Ghz @1.35vlts under 60c in gaming. For just gaming, you don't need any kind of water cooling for a 9900K.

Impressive - how far air coolers have come. What is the fan noise like when system is at full tilt?

 

[Besty]

360 rad for gaming?

My dark rock 4 can keep a 9900K @5Ghz @1.35vlts under 60c in gaming. For just gaming, you don't need any kind of water cooling for a 9900K.

the question was asked by a fellow member who does not want to junk their kit so i answered it.

it depends on your personal sensitivity to noise, i cant use a 360 rad or an air cooler for cooling a 9900k i decided to go triple 580 rad.

 

[Cooper]

What is the fan noise like when system is at full tilt?

Well, its perfectly acceptable to me, but noise levels are a very subjective thing and when I'm gaming I'm always using headphones. The loudest fan noise comes from the GPU, when they go above 70%, it's easily the loudest fan noise coming from the case.

There is this misconception that the 9900K is unusable without expensive AIO's or custom loops. Only if you are heavily into encoding & rendering work, where AVX loads push temps up, do you need to thin about water cooling or if you've shoved a 9900K into a small form factor case, where air flow is severley restricted.

For just desktop work and gaming, a decent high end air cooler is perfectly adequate.

 

[Grim5]

The misconception is the fault of hardware reviewers.

they put cpus under unrealistic loads when measuring temperature

 

[Deleted User 456458]

The misconception is the fault of hardware reviewers.

they put cpus under unrealistic loads when measuring temperature

The reviews are done to cover a wide range of usage scenarios. Gaming, rendering, code crunching, etc.

People use the chip for more than one purpose.

 

[Grim5]

The reviews are done to cover a wide range of usage scenarios. Gaming, rendering, code crunching, etc.

People use the chip for more than one purpose.

Show me one review that shows temperatures for varying use cases?

I'm a gamer, my PC does gaming. That's the only data I want out of the review, I couldn't care less about how the CPU behaves in blender for 10 hours

 

[Deleted User 456458]

Show me one review that shows temperatures for varying use cases?

I'm a gamer, my PC does gaming. That's the only data I want out of the review, I couldn't care less about how the CPU behaves in blender for 10 hours

Then find a gaming loop and run it to where you're stable and feel good about your chip.

There doesn't make the reviewers "faulty" or "Spreading misconception."

 

[Grim5]

Then find a gaming loop and run it to where you're stable and feel good about your chip.

There doesn't make the reviewers "faulty" or "Spreading misconception."

So I have to buy a product to find out how it runs? That doesn't make sense.

I know reviewers are lazy and do worst case testing so that they don't have to test anything else. But it would be nice if they started testing tempreture and power draw in a few scenarios.

 

[Deleted User 456458]

So I have to buy a product to find out how it runs? That doesn't make sense.

Stability testing works like this:

You test across taxing scenarios to determine stability so when a consumer does gaming, they can expect it to run as expected. Different games put on different loads and use different instruction sets. This also changes often thus making using games a useless exercise. You don't know what the next AAA game is going to ask of your CPU so you're always chasing the "gold standard."

Thus you use synthetics that put on consistent loads on a cpu because if you can pass those for a sustained period, your instability in your various games will be due to other areas of the system, not the CPU.

 

[Woodsta888]

Stability testing works like this:

You test across taxing scenarios to determine stability so when a consumer does gaming, they can expect it to run as expected. Different games put on different loads and use different instruction sets. This also changes often thus making using games a useless exercise. You don't know what the next AAA game is going to ask of your CPU so you're always chasing the "gold standard."

Thus you use synthetics that put on consistent loads on a cpu because if you can pass those for a sustained period, your instability in your various games will be due to other areas of the system, not the CPU.

I like this^^ and for your average 'enthusiast', what becomes difficult to decipher with regard to overclocking, is, as a gamer, expect less than those that divulge their spare time in passing a benchmark for long enough to post a score. I think many (including myself in the past) see clocks etc, and benchmark scores and expect those settings to be everyday use. If your are a gamer, to be within 90% of benching scores then, you are doing OK.

 

[Deleted User 456458]

I like this^^ and for your average 'enthusiast', what becomes difficult to decipher with regard to overclocking, is, as a gamer, expect less than those that divulge their spare time in passing a benchmark for long enough to post a score. I think many (including myself in the past) see clocks etc, and benchmark scores and expect those settings to be everyday use. If your are a gamer, to be within 90% of benching scores then, you are doing OK.

Benchmarking and stability testing are colloquially interchanged when they shouldn’t be.

Benchmarking is simply about achieving a score and making sure you don’t regress. Everything is centered around that specific workload and the score derived. Stability has no relevance as long as you can pass the benchmark and up your score.

Stability or stress testing is about putting high stress loads that ensure the component doesn’t fail under varying use cases that it might encounter on a daily basis. That’s why you run p95, OCCT, memtest etc for a reasonable time and just 1 error means it’s not stable.

 

[Steveocee]

Just lightly approached the subject of buying a new CPU with the wife. Of course she asked how much, I said roughly £600 and she offered to buy it me for xmas! Erm.... where's the pre-order button gone again?

 

[Grim5]

Stability testing works like this:

You test across taxing scenarios to determine stability so when a consumer does gaming, they can expect it to run as expected. Different games put on different loads and use different instruction sets. This also changes often thus making using games a useless exercise. You don't know what the next AAA game is going to ask of your CPU so you're always chasing the "gold standard."

Thus you use synthetics that put on consistent loads on a cpu because if you can pass those for a sustained period, your instability in your various games will be due to other areas of the system, not the CPU.

I wasn't aware that reviews were doing stability testing, or that it had anything to do with benchmarking. Thanks for confusing me even more. I was specifically asking for temperature and power draw benchmarks in a small range of tests, not stability testing/hardware destroying software.