Monday, April 20th 2020

Intel Core i7-10700K and i5-10600K Geekbenched, Inch Ahead of 3800X and 3600X

The week has begun with sporadic leaks about Intel's upcoming 10th generation Core "Comet Lake-S" desktop processor family, be it pictures of various socket LGA1200 motherboards, or leaked performance scores. Thai PC enthusiast TUM_APISAK posted links to Geekbench V4 entries of a handful 10th gen Core processors. These include the Core i7-10700K (8-core/16-thread), and the Core i5-10600K (6-core/12-thread). Comparisons with incumbent AMD offerings are inescapable. The i7-10700K locks horns with the Ryzen 7 3800X, while the i5-10600K takes the battle to the Ryzen 5 3600X.

The Core i7-10700K scores 34133 points in the multi-core test, and 5989 in the single-core one. The i5-10600K, on the other hand, puts out 28523 points in the multi-threaded test, and 6081 points in the single-core test. Both scores appear to be a single-digit percentage ahead of the AMD rivals in the multi-threaded test. The Intel chips appear to offer slightly better less-parallelized performance owing to higher boost frequencies for single-threaded or less parallelized workloads. These include an impressive 5.10 GHz max boost frequency for the i7-10700K, and 4.80 GHz for the i5-10600K. APISAK also posted scores of the iGPU-disabled Core i5-10600KF, which is roughly on par with the i5-10600K since it's basically the same chip with its eyes poked out.
Source: TUM_APISAK (Twitter)
Add your own comment

80 Comments on Intel Core i7-10700K and i5-10600K Geekbenched, Inch Ahead of 3800X and 3600X

#26
TheoneandonlyMrK
ppn14nm not competing for server, but for pc we get 9900@ 4,4 157 watts, 3700x @ 4,2 140 watts stress test. Suppose that 10700K@ 4,3 can also be under volted by 0,1 and oc'ed by 200mhz. And AMD had to go for 7nm second gen node n7p to get that. We are yet to witnesses rocket and coVe with great ipc. So 14 nm will kick ass even when amd is on 5Nn.
Kick ass, it'll be kicking mother nature's face more like but yeah if you dream up some unrealistic reality could be a win but no because Intel needs to boost upto 5.3 to get a negligible win Now.
And obviously all clocks and cores equal you get more done with less power on AMD and that simple fact will dent Intel's market share.
Posted on Reply
#28
Dave65
I don't really trust Geekbench much, but the power these things will be drawing is probably going to be incredibly high..... hence the dual optional fire extinguishers...:roll::roll::roll::roll::roll:
Posted on Reply
#29
zlobby
phanbueyThat's great and all but the 8700K OC'd (a ~3 year old chip) is ahead of the 3600x, if these weren't ahead of the ryzen 3 chips with the same core count that would be a massive disaster. The competition is the 4000 series and that's where they're going to have a tough time.
Ahead in what exactly?

I'm pretty sure intel have enough pressure from 3000 Ryzens as it is.
ppn14nm not competing for server, but for pc we get 9900@ 4,4 157 watts, 3700x @ 4,2 140 watts stress test. Suppose that 10700K@ 4,3 can also be under volted by 0,1 and oc'ed by 200mhz. And AMD had to go for 7nm second gen node n7p to get that. We are yet to witnesses rocket and coVe with great ipc. So 14 nm will kick ass even when amd is on 5Nn.
Please don't stop your medication. Stopping it leads to posts like the one above.
Posted on Reply
#30
phanbuey
zlobbyAhead in what exactly?
Are you kidding? In literally just about everything - you name the benchmark an OC'd 8700K will beat a 3600/X... Geekbench, Cinebench, TimeSpy, Games, Apps, SQL... I build and test 3600 rigs all the time for people I know exactly how they handle/bench.

Generally AMD is just now catching up to Intel Coffee Lake (read Kaby, Sky, Sandy etc. etc.) all the same crap at stock w/ Zen 3 core per core -- which is amazing; my point was, that the new chips being ahead of the 3xxx series is not news especially when 4xxx is releasing in 5 months.
Posted on Reply
#31
watzupken
Intel cannot last another generation with their current 14nm and Skylake architecture. They know it and are working to refresh the architecture, though it may still be on 14nm. I would expect this 10xxx series to be the last generation of its kind. Clearly Intel is desperate to push performance to inch ahead of AMD at the expense of power, which they deceptively still show a TDP that is far from reality. I know the argument will be that Intel mentioned that TDP is tied to the base clock, but in order for them to get ahead of competition, they need to engage the boost clock, which drives power way over the TDP. They can get away with the rampant power requirement on the desktop side of things, but not on the laptops. This is clear when the Comet Lake H requires a bigger laptop with beefier cooling, and resulting in a shorter battery life as compared to competition. Whatever refinements they have made to the 14nm fab, they cannot perform magic with it. Something has got to give, and in this case, efficiency. Which is why Intel has been boasting about outperforming competition, but choose to be quiet about the real power requirement to get there.

To get the top end chip here, I would expect an expensive setup since you will need a beefy PSU, motherboard, and high end cooling solution.
phanbueyAre you kidding? In literally just about everything - you name the benchmark an OC'd 8700K will beat a 3600/X... Geekbench, Cinebench, TimeSpy, Games, Apps, SQL... I build and test 3600 rigs all the time for people I know exactly how they handle/bench.

Generally AMD is just now catching up to Intel Coffee Lake (read Kaby, Sky, Sandy etc. etc.) all the same crap at stock w/ Zen 3 core per core -- which is amazing; my point was, that the new chips being ahead of the 3xxx series is not news especially when 4xxx is releasing in 5 months.
I feel AMD's Zen 2 had already outperformed Coffee Lake (Skylake) if you go by the IPC improvements. Clock for clock, I feel AMD should be close if not already faster than Intel. Coffee Lake and also the current Comet Lake is only able to segregate itself from AMD by pushing clockspeed. It is clear that frequency on Skylake was never meant to go this high, since Intel always pride their chips as being energy efficient. As it stands now, energy efficiency is questionable even though they still have the single core performance advantage due to a very high clockspeed. In their desperation to cling on to the performance crown, for the last 4 years they have done nothing other than tweaking the old Skylake chip to make it run at a higher clockspeed.
Posted on Reply
#32
phanbuey
watzupkenI feel AMD's Zen 2 had already outperformed Coffee Lake (Skylake) if you go by the IPC improvements. Clock for clock, I feel AMD should be close if not already faster than Intel. Coffee Lake and also the current Comet Lake is only able to segregate itself from AMD by pushing clockspeed. It is clear that frequency on Skylake was never meant to go this high, since Intel always pride their chips as being energy efficient. As it stands now, energy efficiency is questionable even though they still have the single core performance advantage due to a very high clockspeed. In their desperation to cling on to the performance crown, for the last 4 years they have done nothing other than tweaking the old Skylake chip to make it run at a higher clockspeed.
That was all purely academic, I agree with you regarding your points, but at the end of the day effective OC performance at 24/7 clocks the 8700K (4.9-5.0) was very close to a 2700x in multithreaded apps and crushed it in single thread. If you were going for a high performance all round machine the coffee lakes were faster.

If the 3600x could hit the 4.9-5.0 that the 8700k hits it would absolutely smash it. But it can't so at the end of the day the Coffee Lakes are still faster. These new chips are basically rebranded/tweaked coffee lakes... They will be slightly faster than the 3xxx
Posted on Reply
#33
Melvis
Geekbench?........:slap:

What kind of chilla was used?
Posted on Reply
#34
watzupken
phanbueyThat was all purely academic, I agree with you regarding your points, but at the end of the day effective OC performance at 24/7 clocks the 8700K (4.9-5.0) was very close to a 2700x in multithreaded apps and crushed it in single thread. If you were going for a high performance all round machine the coffee lakes were faster.

If the 3600x could hit the 4.9-5.0 that the 8700k hits it would absolutely smash it. But it can't so at the end of the day the Coffee Lakes are still faster. These new chips are basically rebranded/tweaked coffee lakes... They will be slightly faster than the 3xxx
I don't disagree, but I think for me, I prefer to maintain a balance between a fast and power friendly system. To get the 8700K to close the gap in terms of multithreading performance (against the 2700X), you need to bump the power requirements up significantly. Comparing it with a 3700X, that advantage of OC starts to get eroded on the Intel chip due to the decent IPC gain between Zen+ and Zen 2, not to mention the improved power efficiency.

To summarize, I feel where power is not a problem and user does not mind, Intel will shine because of the higher clockspeed. At this stage however, I tend not to recommend Intel unless there are very specific workload where Intel does well since Intel still charges a sizable premium over AMD, not to mention the additional cost of getting a high end cooler and decent Z series motherboard in order to OC. Gaming used to be a sticking point for AMD and they are still catching up, the gap is now smaller.
Posted on Reply
#35
ZoneDymo
Ex amd fanboyPlaying FC5 on my R5 3600 + 16GB 3733cl16 tuned ballistix spot lt + 5700XT at 1440p full ultra is not a pleasant experience. I am heavily CPU bound. Ryzen is OK for 4K@60Hz gaming, but gaming on a high refresh rate monitor is definitely not the best case scenario for AMD CPUs.
I see you joined yesterday, welcome to the forums man, enjoy your stay.

On your comment/statement, just about every benchmark at 1440p I can find puts the 3600 about 6fps behind an 8700k so yeah while better, you wont notice it.
Posted on Reply
#36
Ex amd fanboy
ZoneDymoI see you joined yesterday, welcome to the forums man, enjoy your stay.

On your comment/statement, just about every benchmark at 1440p I can find puts the 3600 about 6fps behind an 8700k so yeah while better, you wont notice it.
Oh, thank you.

I don`t know what benchmarks you saw to make such a statement. It may be real if 8700k is stock or the ram is not that great. FC5 is really sensitive to single thread perfomance and memory subsystem. Maybe the tested scene was more GPU bound. Idk, but I know that my 5700xt can be around 70% utilization for quiet a long time. Ryzen just can`t give as much frames as gpu can draw. Without being GPU bound 9600k stock is around 10% better than R5 3600 stock (benchmark done by one of the most popular russian tech youtuber. The results are attached with some translation to qualify test setup. I can give a link to this vid, if you want) If both of them are maxed out, 9600k is more than 20% better. So, I would get much more frames with intel cpu in most of the situations.

Posted on Reply
#37
thunderresistant
zlobbyPlease don't stop your medication. Stopping it leads to posts like the one above.
Толсто
watzupkenfor the last 4 years they have done nothing other than tweaking the old Skylake chip to make it run at a higher clockspeed.
They did so since there was 0 rivalry from AMD, so they are still milking the same architecture.
Just business, nothing personal)
Also, if the are feeling real pressure from AMD now they would severely cut their prices, which isn't the case as you know.
So you know whom to blame))
Posted on Reply
#38
Shatun_Bear
You'd need your head examined buying one of these 'last throw of the dice of the 14nm' CPUs. The inefficiency will be a disaster.

Ryzen 4000 will land a few months later with faster single core for like 50% of the power draw and over 40% higher multithreaded perf.
Ex amd fanboyBtw, even 8600k/9600k/9600kf are a better option for gaming, than any Ryzen at the moment. 8700k is WAY faster. That`s why 10th gen is a succes, you will be able to buy a refreshed 8700k for the price of an i5. I don`t like all the synthetic benchmarks tho. They do not represent real world perfomance
Ridiculous FUD. Those CPUs are within a few percent of the 3600 @ 1080p on average, and margin of error or SLOWER @ 1440p and up. And to get these differences you need to be gaming on a 2080 Ti...

And before you get back to me with cherry picked titles unoptimized for Ryzen that show a greater than 5% gaming difference, there are others where the 3600 is faster.
Posted on Reply
#39
the54thvoid
Super Intoxicated Moderator
thunderresistantТолсто


They did so since there was 0 rivalry from AMD, so they are still milking the same architecture.
Just business, nothing personal)
Also, if the are feeling real pressure from AMD now they would severely cut their prices, which isn't the case as you know.
So you know whom to blame))
English only please. I note Google translate derives 'thick'. Hope you're not intending to abuse others?
Posted on Reply
#40
efikkan
Geekbench is known for being unreliable, and generally doesn't translate well to real world performance.

Wait for real benchmarks to judge any product. User submitted benchmarks, even when using good bechmarks, are inherently useless for determining the precise performance of a product, without comparable and fair test conditions. It can be even worse for engineering samples.

Early benchmarks like this are interesting for one thing; hints about how close the next products are. (assuming they are not manipulated, of course)
Posted on Reply
#41
zlobby
I'd really love to see which speculative execution patches are applied to all processors for the benchmarks. Asking for a friend.

Because if you fully patch ANY 9000-series (and previous) intel system it tanks like a rock to the bottom of all benchmarks (no matter how useless the whole benchmarking is). Been there, done that. Such systems feels and perform so sluggish that if it was a horse...

I really love people who only own 1 system for a few years but speak like they have done hands-on research on every piece of hardware out there. Same people who 'play' WinRAR, Blender, etc. all day long.
Posted on Reply
#42
Ex amd fanboy
Shatun_BearYou'd need your head examined buying one of these 'last throw of the dice of the 14nm' CPUs. The inefficiency will be a disaster.

Ryzen 4000 will land a few months later with faster single core for like 50% of the power draw and over 40% higher multithreaded perf.



Ridiculous FUD. Those CPUs are within a few percent of the 3600 @ 1080p on average, and margin of error or SLOWER @ 1440p and up. And to get these differences you need to be gaming on a 2080 Ti...

And before you get back to me with cherry picked titles unoptimized for Ryzen that show a greater than 5% gaming difference, there are others where the 3600 is faster.
Most people have good enough PSUs to handle them. Efficiency is not a big deal.

Waiting is not possible for some people. They want to build here and now. When ryzen 4000 will be on the market and gaming benchmarks will be out, then the discussion will have any sense. But now we can`t say for sure.

Well, maybe you will show some benchmarks that are not cherry picked. I`m really interested, the results you described are surprising
the54thvoidEnglish only please. I note Google translate derives 'thick'. Hope you're not intending to abuse others?
In russian language word "толсто" is sometimes used to say that sarcasm/irony really stands out and can be easily recognized. Idk why he used this word on english speaking forum though
Posted on Reply
#43
cueman
well,not bad from intels 14nm cpu...actually excellent..sure 14nm cpu eat little more power than 7nm amd cpu. but hey, amd cpu cant oc so much,its near top... example 3700x not go over 4400mhz.

looks amd time end closing fast, anyway,i wait intels10nm and finally 7nm cpus, then we see real compare,when both have even tech.


both should coming this year.
Posted on Reply
#44
John Naylor
CPUs are tools ... Unless my job is to run geekbench all day, I don't see a value to the results. Manufacturers love to show benchmarks in which their product excels.... so they pick ones in whch theirproduct is shown in a favorable light and then the brand loyalists will post these on technical forums. I have so many people submit builds and when I ask why they chose a particular CPU, I get ...
  • It has [insert number] nm die size
  • It has [insert number] cores
  • It has [insert number] Ghz
  • It has [insert number] whatever
  • It scored [insert number] in this [irrelevant] benchmark
And when asked how it performs in the apps / games they actually run, they have no idea. Most builds we are asked about are in the $500 3900X or 9900KF category. Whiuch one is better ? I can't answer hat question.... i can only answer which is better for the apps you use but 1st you have to tell me what they are.

Gaming, CAD, Photo Editing, Video Editing and Gaming,Sound encoding, the better choice is Intel
Rendering, Brain Simulation, Software Development, H264/H265 Encoding, the better choice is AMD

Geekbench doesn't give the answers I need to make a valid comparison.
Posted on Reply
#45
Shatun_Bear
Ex amd fanboyMost people have good enough PSUs to handle them. Efficiency is not a big deal.

Waiting is not possible for some people. They want to build here and now. When ryzen 4000 will be on the market and gaming benchmarks will be out, then the discussion will have any sense. But now we can`t say for sure.

Well, maybe you will show some benchmarks that are not cherry picked. I`m really interested, the results you described are surprising


In russian language word "толсто" is sometimes used to say that sarcasm/irony really stands out and can be easily recognized. Idk why he used this word on english speaking forum though
If you have to buy in the 'here and now', Ryzen CPUs cover all use cases as the prices are extremely good value for money.

Starting at the low end with the 6-core £80 1600AF all the way up to the 12-core 3900X for £400. Same with the motherboards across 3 generations from low end to high end, choose which suits you. This is the beauty of AM4's cross generational compatability.
Posted on Reply
#46
zlobby
Shatun_BearIf you have to buy in the 'here and now', Ryzen CPUs cover all use cases as the prices are extremely good value for money.

Starting at the low end with the 6-core £80 1600AF all the way up to the 12-core 3900X for £400. Same with the motherboards across 3 generations from low end to high end, choose which suits you. This is the beauty of AM4's cross generational compatability.
Although AMD mobos still don't get the same love as intel's, even though there is a huge improvement in the recent years, particularly after Ryzen initially launched.
Posted on Reply
#47
Ex amd fanboy
Shatun_BearIf you have to buy in the 'here and now', Ryzen CPUs cover all use cases as the prices are extremely good value for money.

Starting at the low end with the 6-core £80 1600AF all the way up to the 12-core 3900X for £400. Same with the motherboards across 3 generations from low end to high end, choose which suits you. This is the beauty of AM4's cross generational compatability.
Ryzen CPUs do not cover all use cases (high refresh rate, low latency gaming), but I can`t argue that cross generational compatability is a great feature. Zen 2 is also a good "plug n play". Turn on XMP, PBO and you are good to go. But intel is currently the only option for the highest frame rates possible.
zlobbyAlthough AMD mobos still don't get the same love as intel's, even though there is a huge improvement in the recent years, particularly after Ryzen initially launched.
What do you mean by "love"?
Posted on Reply
#48
hurakura
and where are the Ryzen scores for comparison? Those Intel scores are actually lower that those of Ryzen 5 and 7. If you check GeekBench site you can see lots of Ryzen 7 3800x scores a lot higher than those of Intel's I7 10700K.

Other sites publishing this news are at least putting Ryzen results as comparison, but even then they look for the lowest Ryzen score on GeekBench.
Intel is only strong in the marketing department and paying sites to publish fabricated news.
Posted on Reply
#50
zlobby
phanbueyGames, Apps, SQL...
You kind of lost me there.
Posted on Reply
Add your own comment
Nov 5th, 2024 02:24 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts