First Core i7-9700K 'review' hits web; more Core i9-9900K and Core i5-9600K benchmarks...

midian182

Posts: 9,722   +121
Staff member
Highly anticipated: After Intel’s upcoming Core i7-9700K appeared on Geekbench last week, a full review of the 9th-gen chip has revealed more benchmark scores. We've also seen new numbers for the Core i9-9900K and Core i5-9600K.

The previous Geekbench score showed the Core i7-9700K gaining a 6297 single-core score and 30152 in multi-core performance. While this marked a roughly 20 percent improvement over the Ryzen 2700X, it was just 4 percent better than the 8700K.

Spanish-language website El Chapuzas Informatico has what is alleged to be 'the first Core i7-9700K review.' It confirms that it is a 14nm CPU without hyper-threading (8-core/8-thread) and has a 4.9GHz single-core boost clock and 4.6GHz boost across all cores.

The benchmarks were run using an upcoming Z390 motherboard, MSI’s GeForce GTX 1070 Gaming Z, and an unspecified amount of G.Skill TridentZ DDR4-3600 RAM. While Intel’s chip came top in many of the results—it was the fastest in all the gaming tests—there were some benchmarks where it was beaten by AMD’s Ryzen 2700X. There are also many cases where the 9700K shows only a marginal improvement over the 8700K.

The publication was able to overclock the chip to 5GHz (all cores) on 1.4 Volts. In late August, leaks showed it reaching 5.5GHz on all eight cores using water cooling with the voltage set at 1.536V on a Z370 mobo.

As always, it’s best to take all this with a pinch of salt. It’s better to wait until the Coffee Lake Refresh chips arrive, supposedly next month, to see more reviews and draw solid conclusions. But the 9700K, which is expected to cost around $400, could become the CPU of choice for gamers. Check out all the benchmarks here.

In other Coffee Lake Refresh news, Geekbench numbers for the Core i9-9900K and Core i5-9600K have leaked. The high-end 8-core/16-thread chip managed 6248 points in single-core and 33,037 points in multi-core benchmarks, while the 6-core/6-thread CPU hit 6,027 and 23,472 in the same tests. The current Core i7-8700K flagship, for comparison, gained 5,508 single-core and 25,034 multi-core results. These were achieved using a Z370 mobo, so the numbers could improve when running on the Z390 platform.

Permalink to story.

 
I was looking at the I-7 6700K performance compared to newer generation Intel chips and they really haven't made any improvements. The synthetic benchmarks looked a lot better, but in the real world benchmarks they performed almost identical. I have a 6700K and feel like I got a good value, considering they really haven't changed anything significant since then. Intel is just milking its products for as long as it can, shitty company.
 
I was looking at the I-7 6700K performance compared to newer generation Intel chips and they really haven't made any improvements. The synthetic benchmarks looked a lot better, but in the real world benchmarks they performed almost identical. I have a 6700K and feel like I got a good value, considering they really haven't changed anything significant since then. Intel is just milking its products for as long as it can, shitty company.

Also they actually don't have new architecture as a reply to Ryzen, which really caught them with their pants down!!!
What is baffling is they are still on that 5%-10% improvement by mix of core count/ hyper threading.
 
Also they actually don't have new architecture as a reply to Ryzen, which really caught them with their pants down!!!
What is baffling is they are still on that 5%-10% improvement by mix of core count/ hyper threading.
To be fair, they did slap on 2 extra cores on the newer chips, but it obviously isn't helping them according to the benchmarks (real world and synthetic). Everyone used to say AMD were the chips that ran hot and used a lot of power, now that has flipped after Intel slapped on the two extra cores and raised the clock speeds on the same chip. If I were in the need to build a PC today, it would definitely be with a Ryzen CPU.
 
I was looking at the I-7 6700K performance compared to newer generation Intel chips and they really haven't made any improvements. The synthetic benchmarks looked a lot better, but in the real world benchmarks they performed almost identical. I have a 6700K and feel like I got a good value, considering they really haven't changed anything significant since then. Intel is just milking its products for as long as it can, shitty company.

Also they actually don't have new architecture as a reply to Ryzen, which really caught them with their pants down!!!
What is baffling is they are still on that 5%-10% improvement by mix of core count/ hyper threading.

The problem is Intel really doesn't have much of a way to make it's CPUs faster. They can't add more cores without violating thermal constraints, and they can't improve the architecture much more then it is.

Long term, Intel needs to retire Core. Its simply too long in the tooth at this point.
 
Those extra cores dont seem to fix the big issue, that is per clock this is no better then skylake.

Intel cant staple more cores on with the efficiency that AMD can, and if Zen 2 comes with a 5-10% IPC increase, intel's only advantage, now that power consumption is thrown to the birds and adding more cores isnt helping them, is higher clock rates, and that kind of advantage is tenuous.

I've got a ryzen 1700 on the way, and am eagerly waiting for the zen 2 release next year. It seems intel is SOL until their 10nm node is ready, and even then they have to do something about core's lack of IPC improvements, without relying on cache tricks that the next meltdown/spectre will annihilate performance wise.
Aaanndd the war begins
The time of intel domination is over; Begun, the core war has.
 
Intel definitely needs new architecture in a chip and lets not be naive, I'm sure they have working on it for years. Let's also not get carried away with ryzen 2. Faboys screamed the ryzen refresher would be 5-10% increase in IPC, it turned out to be 2%. Ryzen 2 will not be an increase of 5-10% either, AMD will be more then lucky if they hit 5% and will most likely hit half that mark. IPC gains are far more difficult to achieve then fanboys understand as both Intel and AMD are experiencing those difficulties.
 
Intel definitely needs new architecture in a chip and lets not be naive, I'm sure they have working on it for years. Let's also not get carried away with ryzen 2. Faboys screamed the ryzen refresher would be 5-10% increase in IPC, it turned out to be 2%. Ryzen 2 will not be an increase of 5-10% either, AMD will be more then lucky if they hit 5% and will most likely hit half that mark. IPC gains are far more difficult to achieve then fanboys understand as both Intel and AMD are experiencing those difficulties.
To be fair, any fanbois screaming about a 5% increase with zen + were rather dumb. Zen+ was just zen on 12nm, with cache and IMC improvements. Even then, AMD managed a 3-4% increase IIRC, which is still impressive.

Zen 2 is supposed to be an actual architectural improvement, and thus 5-10% expectations are not without merit, as we know AMD stil.l has a lot of low hanging fruit with zen's memory controller, infinity fabric latency, cache latency, ece. AMD also has a lot of room to work on clock speeds.
 
It's more like AMD has a lot of work to improve speeds Anandtech clocked Zen + at 3% IPC increase and for gaming it was a 1-2% clock for clock improvement. AMD fan boys are going to be disappointed with Zen 2.
 
Intel definitely needs new architecture in a chip and lets not be naive, I'm sure they have working on it for years.
Well I highly doubt that; I think after the Bulldozer fiasco they were confident that AMD cannot pull one over their heads but they did, otherwise they would not have gone into the hiring spree (Jim* esp.) just after it and would have released the "new" architecture by now.
 
Well I highly doubt that; I think after the Bulldozer fiasco they were confident that AMD cannot pull one over their heads but they did, otherwise they would not have gone into the hiring spree (Jim* esp.) just after it and would have released the "new" architecture by now.

That's a nice thought but in the real world tech companies spend years and millions of dollars in research & development that are in various stages. Intel and AMD are no different in that department. I'm sure AMD has started the ground work for architecture after Ryzen 2.
 
Those extra cores dont seem to fix the big issue, that is per clock this is no better then skylake.

Intel cant staple more cores on with the efficiency that AMD can, and if Zen 2 comes with a 5-10% IPC increase, intel's only advantage, now that power consumption is thrown to the birds and adding more cores isnt helping them, is higher clock rates, and that kind of advantage is tenuous.

I've got a ryzen 1700 on the way, and am eagerly waiting for the zen 2 release next year. It seems intel is SOL until their 10nm node is ready, and even then they have to do something about core's lack of IPC improvements, without relying on cache tricks that the next meltdown/spectre will annihilate performance wise.

The time of intel domination is over; Begun, the core war has.
nice
 
Intel definitely needs new architecture in a chip and lets not be naive, I'm sure they have working on it for years. Let's also not get carried away with ryzen 2. Faboys screamed the ryzen refresher would be 5-10% increase in IPC, it turned out to be 2%. Ryzen 2 will not be an increase of 5-10% either, AMD will be more then lucky if they hit 5% and will most likely hit half that mark. IPC gains are far more difficult to achieve then fanboys understand as both Intel and AMD are experiencing those difficulties.

Ryzen Refresh was never supposed to be anything else than slightly bigger clocks. And architecture is exactly same that is used on Ryzen.

It's more like AMD has a lot of work to improve speeds Anandtech clocked Zen + at 3% IPC increase and for gaming it was a 1-2% clock for clock improvement. AMD fan boys are going to be disappointed with Zen 2.

+3% IPC increase without making any changes to architecture is actually pretty good one.
 
Back