MetroidDo you really believe the 9900k consumes only 18 watts more than the 9700k, 100% multithreading?

Yes I do. Because of the variance between chips. If the 9700K was good enough to be a 9900K, it would be. But because it can't, it slides down a bin and drops HT. HT yields a few C increase in temps so I wouldn't doubt the validity in THAT test. Now, you are banging on your unrelated drum about a freakin stress test which isn't a real world situation in the first place. You go with your bad self.

EarthDogYes I do. Because of the variance between chips. If the 9700K was good enough to be a 9900K, it would be. But because it can't, it slides down a bin and drops HT. HT yields a few C increase in temps so I wouldn't doubt the validity in THAT test. Now, you are banging on your unrelated drum about a freakin stress test which isn't a real world situation in the first place. You go with your bad self.

I would agree with you if the 9900k had more cores and had to be a better bin, difference is only 100mhz between them and price wise like I said intel charges more for the 9900k because of the hyperthreading, not the bin itself, there are 9700k that uses much less watts than the 9700k in that review, so your point about the 9900k to have a better bin is irrelevant and by the way if I had a 9700k, it would consume much less watts than in that review but i would not say much about the 9900k because of the hyperthreading on it.

MetroidI would agree with you if the 9900k had more cores and had to be a better bin, difference is only 100mhz between them and price wise like I said intel charges more for the 9900k because of the hyperthreading, not the bin itself, there are 9700k that uses much less watts than the 9700k in that review, so your point about the 9900k to have a better bin is irrelevant and by the way if I had a 9700k, it would consume much less watts than in that review but i would not say much about the 9900k because of the hyperthreading on it.

Longest. (run on) Sentence. EVER. :p


The bin and the HT fetch more money, yes. There are bins that go either way, some leaky, some not so results will vary between CPUs. The 9900K is a better binned CPU than a 9700K in most cases. There will always be exceptions as no two CPUs are the same. If you disable HT on the 9900K it will look more like a 9700K....................................................because that's what it is (for all intents and purposes).

EarthDogLongest. (run on) Sentence. EVER. :p


The bin and the HT fetch more money, yes. There are bins that go either way, some leaky, some not so results will vary between CPUs. The 9900K is a better binned CPU than a 9700K in most cases. There will always be exceptions as no two CPUs are the same. If you disable HT on the 9900K it will look more like a 9700K....................................................because that's what it is (for all intents and purposes).

Lets forget the bin, take a 9900k and test yourself with and without hyperthreading and then you will get the result on how much power the hyperthreading uses like I did with my i7 920 few years ago. There are no reviews on that, I wish it had. I believe that 9900k would use much less volts without hyperthreading.

MetroidLets forget the bin, take a 9900k and test yourself with and without hyperthreading and then you will get the result on how much power the hyperthreading uses like I did with my i7 920 few years ago. There are no reviews on that, I wish it had. I believe that 9900k would use much less volts without hyperthreading.

Ive done the testing... for generations. I'm also (was) an extreme overclocker so I know how to play with my chips with and without HT and what (generally) happens.

Of course the 9900k (any CPU with HT) can use a bit less volts with (much seems a bit, well, much, lol) HT disabled... this is common knowledge. There are many factors which go into the power results ;).

Point is, I fully believe those results listed in real world applications and could give two hoots about response in a stress testing application's power use considering most are above and beyond real world activities. Im sure it uses a lot more power in stress tests... nobody would disagree. That said, I believe those results with real world apps. I've literally done that kind of testing for years.

Metroidtest it on intelburntest and see yourself your house burn together hehe

Please do not confuse temperature in degrees with the wattage its putting out. Ive burned my finger on a 5W IC in a built-for-purpose mining rig... ;)

Remember, a bonfire with yellow flames is just as hot temperature wise as a lighter with a yellow flame... but which has more energy behind it?

Intel will only keep the one fab on 10nm and have reverted the others back to 14. They will lose money on every chip sold probably for the life of 10nm lol. It's only in production to say they have it. Purely, a technicality. Incredibly embarrassing.

EarthDogIve done the testing... for generations. I'm also (was) an extreme overclocker so I know how to play with my chips with and without HT and what (generally) happens.

Of course the 9900k (any CPU with HT) can use a bit less volts with (much seems a bit, well, much, lol) HT disabled... this is common knowledge. There are many factors which go into the power results ;).

I said my i7 920 used 1.18v hyperthreading off and 1.40v when was on, manual overclocking and as i stated it used more than half on stress tests. So if you are " an extreme overclocker " then I really dont understand what you tried to imply. What I said is 18 watts only on the 9900k more x 9700k i said that while might be correct stock wise and different bins, i implied that is not entirely true in most cases.

MetroidI said my i7 920 used 1.18v hyperthreading off and 1.40v when was on, manual overclocking and as i stated it used more than half on stress tests. So if you are " an extreme overclocker " then I really dont understand what you tried to imply. What I said is 18 watts only on the 9900k more x 9700k i said that while might be correct stock wise and different bins, i implied that is not entirely true in most cases.

Reviews say 30 watts.

TheGuruStudReviews say 30 watts.

review here at tpu says 18 watts in multithread tests but yeah I agree with you, 30 or more watts is the normal x 9700k, here at tpu might have had a better bin on stock voltage.



review here at tpu says 18 watts in multithread tests but yeah I agree with you, 30 or more watts is the normal x 9700k, here at tpu might have had a better bin on stock voltage.




Here on anandtech says a lot more, 44 watts.

MetroidI said my i7 920 used 1.18v hyperthreading off and 1.40v when was on, manual overclocking and as i stated it used more than half on stress tests. So if you are " an extreme overclocker " then I really dont understand what you tried to imply. What I said is 18 watts only on the 9900k more x 9700k i said that while might be correct stock wise and different bins, i implied that is not entirely true in most cases.

If that is actually true (....) that would be the most voltage I have ever seen between HT enabled and disabled for the same clock speed. Typically its around .05-/1V in my experience and/or yields 100-300 Mhz more at the same voltage. I don't ever recall there being a .2V difference just for HT in any of the chips I have done that with, including an i7 920.

Anyway, just know its believable, that result you think isn't...mmkay? There are different results from different tests. :)

The Anandtech power is also a stress testing application, btw (P95).

EarthDogIf that is actually true (....) that would be the most voltage I have ever seen between HT enabled and disabled for the same clock speed. Typically its around .05-/1V in my experience and/or yields 100-300 Mhz more at the same voltage. I don't ever recall there being a .2V difference just for HT in any of the chips I have done that with, including an i7 920.

Anyway, just know its believable, that result you think isn't...mmkay? :)

Well that is how it was with my chip, not sure if i can find the screenshot, I posted on xtremesystems in 2009. I have been aside from xtreme overclocking so I have no idea how is today.

Well, data previous than 2010 cant be found, I found this, many images could not be loaded, it has been a long time.

www.xtremesystems.org/forums/showthread.php?205779-Core-i7-X58-Overclocking-Thread&p=3680338&viewfull=1#post3680338
www.xtremesystems.org/forums/showthread.php?261294-Could-you-pros-check-a-n00bs-work&p=4601961&highlight=#post4601961
www.xtremesystems.org/forums/showthread.php?205779-Core-i7-X58-Overclocking-Thread&p=4588997&viewfull=1#post4588997

These images could be loaded, the others no.



Anyway, I tried but you get my point.

EarthDogIf that is actually true (....) that would be the most voltage I have ever seen between HT enabled and disabled for the same clock speed. Typically its around .05-/1V in my experience and/or yields 100-300 Mhz more at the same voltage. I don't ever recall there being a .2V difference just for HT in any of the chips I have done that with, including an i7 920.

Anyway, just know its believable, that result you think isn't...mmkay? There are different results from different tests. :)

The Anandtech power is also a stress testing application, btw (P95).

They use POV-Ray. In P95 its closer to 200w and with AVX its 250w when OC at 5ghz

AnandtechFor our testing, we use POV-Ray as our load generator then take the register values for CPU power.

Unless the board settings sticks to spec its going to suck up juice.

That must be an updated test suite (or the CPU tesrs?). P95 is what I recall using on z370 when writing for them. Thanks for the clarification. :)

> These will also feature DLBoost, a fixed-function maxtrix-multiplication hardware that
> speeds up deep-neural net building and training by 5x, as well as certain AVX-512 instructions.

then

> The first of these chips will target mobile computing platforms

rip desktop users, none of that sounds interesting.

ncrsThis is not the first time they put out a broken 10nm product just as a smoke screen for investors. The first one was i3-8121U - a dual-core processor that was so broken it had to have the iGPU disabled. According to rumors the yields for this model were below 1%. It cost Intel millions to make at most thousands of "full functioning" ones.

You're telling us what the majority of TPU users already know and you're just repeating old news. The rest is just dishonesty and rumor, on your part, in an effort to support a fellow member of your "side".

MetroidI said my i7 920 used 1.18v hyperthreading off and 1.40v when was on, manual overclocking and as i stated it used more than half on stress tests. So if you are " an extreme overclocker " then I really dont understand what you tried to imply. What I said is 18 watts only on the 9900k more x 9700k i said that while might be correct stock wise and different bins, i implied that is not entirely true in most cases.

I’m confused as to what your tests on the thoroughly ancient i7-920 have to do with disproving as you say, the 18 watt difference between the 9700k and 9900k.

rtwjunkieI’m confused as to what your tests on the thoroughly ancient i7-920 have to do with disproving as you say, the 18 watt difference between the 9700k and 9900k.

I pointed the 18 watts is not real, it's a lot more than that. I do respect the test procedure though but that is not the reality of how much power consumption hyperthreading uses, comparing 9700k x 9900k is not right as well, just test the 9900k with and without hyperthreading and you will have an answer like i did when had the opportunity few years ago. I do not have a cpu with hyperthreading here, so I cant test it. Basically in my view 9900k is a failed product because it has hyperthreading and lots of cores, too much heat, is time for amd follow intel and give the choice if you want hyperthreading/smt or not, so users can decide about it and that is why I think the 9700k is a step in the right direction, there are enough cores already, most people dont need hyperthreading anymore and charging $100 for it is too much, however only you can decide if is a good deal, in my case i dont see, everything works better with hyperthreading off or not having hyperthreading in my applications.

If you are an overclocker you can reduce the power consumption a lot by reducing cpu vcore but the catch is you will need to disable hyperthreading and if you disable hyperthreading on a 9900k then why haven't bought a 9700k then? see, okay so 9900k might have a better bin, so disable ht and enjoy, hyperthreading uses too much electricity and it gets worse if people dont have an adequate cooling solution, at that time i had an amazing watercooling solution, custom made with 2 pumps 18 watts each, triple radiator. Was it worth? at that time it was hehe but today I see is not worth for me. I rather have something that can be used on air cooling and cpus with tdp lower than 95 watts is good to go, my i7 920 was130 watts but 130 watts was at 2.4 ghz hehe, at 4.2ghz used around 350watts.

how do they come up with these names "sunny cove" "icy lake" and so on

r.h.phow do they come up with these names "sunny cove" "icy lake" and so on

Well, they were sailing on their 10 million dollar yachts and saw a nice cove in the sun. Then they were sitting in their 5 mil vacation home in the mountains next to a lake.

They spent so much time using 100s to light cigars that the retards forgot how to make CPUs. Now, they're going to burn in hell and make VIA look good.

Whiskey lake should be pretty self-explanatory.

Can we expect a TPU review of the case badge, including overclocking results and temperature results with a few popular coolers? Maybe even some delidding?

Manu_PTIntel is not worried at all. It would be as simple as dropping prices. Imagine dropping 9900k to 400€ and 9700k to 300€. They dont do it because they dont need to. They have the best performance.

Of course they're not worried - the incompetent and bullish ones never are. Until it's too late that is. Nokia was also not worried btw and we all know how that went.
Intel will NEVER drop their prices. That's wishful fanboy dreaming.

TomorrowIntel will NEVER drop their prices. That's wishful fanboy dreaming.

Want to bet?

MetroidI pointed the 18 watts is not real, it's a lot more than that. I do respect the test procedure though but that is not the reality of how much power consumption hyperthreading uses, comparing 9700k x 9900k is not right as well, just test the 9900k with and without hyperthreading and you will have an answer like i did when had the opportunity few years ago. I do not have a cpu with hyperthreading here, so I cant test it. Basically in my view 9900k is a failed product because it has hyperthreading and lots of cores, too much heat, is time for amd follow intel and give the choice if you want hyperthreading/smt or not, so users can decide about it and that is why I think the 9700k is a step in the right direction, there are enough cores already, most people dont need hyperthreading anymore and charging $100 for it is too much, however only you can decide if is a good deal, in my case i dont see, everything works better with hyperthreading off or not having hyperthreading in my applications.

If you are an overclocker you can reduce the power consumption a lot by reducing cpu vcore but the catch is you will need to disable hyperthreading and if you disable hyperthreading on a 9900k then why haven't bought a 9700k then? see, okay so 9900k might have a better bin, so disable ht and enjoy, hyperthreading uses too much electricity and it gets worse if people dont have an adequate cooling solution, at that time i had an amazing watercooling solution, custom made with 2 pumps 18 watts each, triple radiator. Was it worth? at that time it was hehe but today I see is not worth for me. I rather have something that can be used on air cooling and cpus with tdp lower than 95 watts is good to go, my i7 920 was130 watts but 130 watts was at 2.4 ghz hehe, at 4.2ghz used around 350watts.

I'd test this, but only have a 16c/32t CPU. The last time I did this, the 8800k could go up 200 hz or I could lower voltage 0.7xV.

There are several generations between your testing (which I never saw back then) and today's reality. :)

9900k is not a failure, btw. Not even close.

Anyway..OT. I digress. :)

On Servers parts they do drop prices, they have no other choice since 32C EPYCS are around + all the security problems, this is why we gonna see their margins getting lower every quarter.
www.servethehome.com/intel-is-serving-major-xeon-discounts-to-combat-amd-epyc/

TheGuruStudWhil
Intel will only keep the one fab on 10nm and have reverted the others back to 14. They will lose money on every chip sold probably for the life of 10nm lol. It's only in production to say they have it. Purely, a technicality. Incredibly embarrassing.

10nm and 14nm doesn't share production lines, and share very little equipment.
Intel have reserved more 14nm production capacity for CPUs than ever before, and will not go full-scale on 10nm, but they are not "reverting" production lines back to 14nm.

efikkan10nm and 14nm doesn't share production lines, and share very little equipment.
Intel have reserved more 14nm production capacity for CPUs than ever before, and will not go full-scale on 10nm, but they are not "reverting" production lines back to 14nm.

I read they were in the middle of upgrading then abandoned it. This aligns with another article I read Dec of last year saying they were going to use two fabs for 10nm, but we know that's not happening.