• Happy holidays, folks! Thanks to each and every one of you for being part of the Tom's Hardware community!

News Intel 10th Gen Comet Lake Desktop CPUs Rumored to be Announced On April 30

Is this a reliable source or this site beoming the new wccftech?

Because in the article "Approach the rumored dates with a bit of caution since we have no way to confirm them. ".
It is not like we don't know everthing about this "new" excitnig skylake cpu already.

Lga 1200, 14nm++++++++++, skylake architecture, same performance with slightly better boost.

There is no need for rumors. Everyone knows how skylake perform because it is in the market for 3 years.

If this is the exact release date so I am happy to know but it does not sound like that from the article.
 
  • Like
Reactions: TCA_ChinChin
Is this a reliable source or this site beoming the new wccftech?

Because in the article "Approach the rumored dates with a bit of caution since we have no way to confirm them. ".
It is not like we don't know everthing about this "new" excitnig skylake cpu already.

Lga 1200, 14nm++++++++++, skylake architecture, same performance with slightly better boost.

There is no need for rumors. Everyone knows how skylake perform because it is in the market for 3 years.

If this is the exact release date so I am happy to know but it does not sound like that from the article.
You can add as many + as you want, kid - this will still outsell everything AMD has. Funny how you deride Skylake - yet AMD has yet to reach parity with that ancient 3 year old architecture....
 
You can add as many + as you want, kid - this will still outsell everything AMD has. Funny how you deride Skylake - yet AMD has yet to reach parity with that ancient 3 year old architecture....
It will perform better.... we all know it because of the 9900ks (special edition that no one can find) benchmarks.
I asked if this rumor is even slightly correct, troll.
Leave the guessing to wccftech.
 
You can add as many + as you want, kid - this will still outsell everything AMD has.
On the pre-built market where most buyers are basically going for updated versions of whatever they already had, sure. In the DIY space though, hardware sites and channels that periodically comment on things like results from Amazon affiliate links say AMD is out-selling Intel by 2-3X, which is in line with what Amazon rankings and sales stats from the few online stores that publish those also report.

Intel may still have the crown for highest attainable frame rates at any cost but most people who aren't buying blind want more practical and cost-effective CPUs.
 
You can add as many + as you want, kid - this will still outsell everything AMD has. Funny how you deride Skylake - yet AMD has yet to reach parity with that ancient 3 year old architecture....
AMD has surpassed even Coffee-lake in everything except AVX in certain scenarios and FPS in certain scenarios. If those are the only things that matter, then I won't disagree.
 
You can add as many + as you want, kid - this will still outsell everything AMD has. Funny how you deride Skylake - yet AMD has yet to reach parity with that ancient 3 year old architecture....

If early benchmarks are to be believed... Not really. Single core slightly higher, and multi core still doesn't beat the 3900x. But more expensive because of the additional cores, and running hotter than the 9900k.

https://browser.geekbench.com/processors/2587

https://browser.geekbench.com/v5/cpu/1584149
 
On the pre-built market where most buyers are basically going for updated versions of whatever they already had, sure. In the DIY space though, hardware sites and channels that periodically comment on things like results from Amazon affiliate links say AMD is out-selling Intel by 2-3X, which is in line with what Amazon rankings and sales stats from the few online stores that publish those also report.
Look in 2015 when AMD announced that ZEN was coming up AMD's revenue dropped like a stone, which is understandable because why would you buy bulldozer if ZEN is on the horizon, so whoever is buying ZEN is also just going for updated versions of whatever they already had.
AMDs gross revenue is only slightly above 2010-2012 levels.
https://www.macrotrends.net/stocks/charts/AMD/amd/revenue
While their net income is lower than what it was back then.
I guess ZEN isn't quite as cheap to make as people like to think.
https://www.macrotrends.net/stocks/charts/AMD/amd/net-income
 
Given that the launch of these processors was scheduled for March 30th, postponing announcement for exactly one month and launch for 6 weeks sounds plausible. However, I doubt the situation with the pandemic will get better by then, especially in the US where it is only just starting – if anything for many states the situation is expected to peak by then, unless a 6-week lockdown is enforced nationwide.

AMD has surpassed even Coffee-lake in everything except AVX in certain scenarios and FPS in certain scenarios. If those are the only things that matter, then I won't disagree.
That’s certainly not true. First of all IPC is a per workload metric and the idea that AMD has superior IPC to Intel in every non-AVX workload is for the birds. Also, even if AMD cpus did have such a total IPC superiority (which they don’t) it is only a minor one and is more than offset by Intel’s clock speed advantage. So Intel still holds the single-threaded performance crown in most workloads and anything scalable only up to the number of cores/threads Intel is offering (hence quad-threaded, hexa-threaded, eight-threaded performance especially in an OCed vs OCed comparison).

It should also be noted that AMD’s cpus even if they did have an IPC superiority is not something that should be attributed to the sheer brilliance of their microarchitecture. Zen2 is using double the amount of per core L2 cache (512KB Vs 256KB) and double the amount of per core L3 cache (4MB vs 2MB) compared to Skylake. Substantially increasing the capacity of your CPU memory subsystem is only a trick you can do once per 5-10 years. AMD is running out of cheap tricks and is time to actually improve the computationally efficiency of its core architecture than just relying on throwing more cores and more cache in order to mitigate the inherent deficiencies of its current design.

And by the way, we hear so much about Intel cpus being supposedly insecure against exotic side-channel attacks but the ones who actually get actually badly hacked in the real world (and remotely) are AMD themselves and that with regular attacks! Not exactly confidence inspiring when a company selling cpus claiming to be secure has such a major hack and has their IP Verilog files stolen under their nose.
 
Looking forward to seeing what the 10700K and 10900K bring to the gaming performance table!
Spoiler: it won't be anything that you haven't already seen with the 9900k, same old Skylake, only slightly faster clocks and up to two extra cores.

I'd be far more interested in seeing how Zen 3 will turn out: more uArch improvements, all cores sharing a single larger L3 cache which eliminates the tier of fabric between CCXes and the CCD-IO interface and should reduce latencies all around. Zen 2 brought AMD roughly on par with Intel's best in the mainstream, Zen 3 will likely give it a definitive lead at least until Intel launches its next post-Skylake desktop architecture on 10nm++ and beyond.
 
Given that the launch of these processors was scheduled for March 30th, postponing announcement for exactly one month and launch for 6 weeks sounds plausible. However, I doubt the situation with the pandemic will get better by then, especially in the US where it is only just starting – if anything for many states the situation is expected to peak by then, unless a 6-week lockdown is enforced nationwide.


That’s certainly not true. First of all IPC is a per workload metric and the idea that AMD has superior IPC to Intel in every non-AVX workload is for the birds. Also, even if AMD cpus did have such a total IPC superiority (which they don’t) it is only a minor one and is more than offset by Intel’s clock speed advantage. So Intel still holds the single-threaded performance crown in most workloads and anything scalable only up to the number of cores/threads Intel is offering (hence quad-threaded, hexa-threaded, eight-threaded performance especially in an OCed vs OCed comparison).

It should also be noted that AMD’s cpus even if they did have an IPC superiority is not something that should be attributed to the sheer brilliance of their microarchitecture. Zen2 is using double the amount of per core L2 cache (512KB Vs 256KB) and double the amount of per core L3 cache (4MB vs 2MB) compared to Skylake. Substantially increasing the capacity of your CPU memory subsystem is only a trick you can do once per 5-10 years. AMD is running out of cheap tricks and is time to actually improve the computationally efficiency of its core architecture than just relying on throwing more cores and more cache in order to mitigate the inherent deficiencies of its current design.

And by the way, we hear so much about Intel cpus being supposedly insecure against exotic side-channel attacks but the ones who actually get actually badly hacked in the real world (and remotely) are AMD themselves and that with regular attacks! Not exactly confidence inspiring when a company selling cpus claiming to be secure has such a major hack and has their IP Verilog files stolen under their nose.
Give me some examples of Intel offering better performance than AMD in non-AVX scenarios or scenarios where Intel's lower memory latency gives them better performance. I can see areas such as Solidworks and Adobe suite being Intel advantages (so some degree), due to companies spending more than a decade optimizing exclusively for the Intel Core architecture. The furthermore, for productivity workloads, one doesn't compare AMD vs Intel in a equal core count scenario, as the whole point of AMD is that it has more cores for cheaper.

Also, increasing L2 and L3 cache is a GOOD THING. Even if the huge L3 cache is partly to make up for their increased memory latency, I can't see how that is a "cheap trick". Anything that gets a performance advantage is a "cheap trick" then? Is Nvidia DLSS 2.0 or RTX a "cheap trick"? Is Intel adding clock speeds a "cheap trick"?

I didn't say anything about exotic side-channel attacks, but since you're the one who brought it up, but Intel are the only ones that seem to have so many CVEs to their name. AMD stolen IP isn't relevant to their current and future technology and bringing up something like that when talking about CPU performance simply shows the amount people have to stretch their imaginations to delude themselves that Intel is still currently relevant.