Now the question is what AMD did to solve the latency problem of needing multiple hops to reach system memory. Or whether they solved it. Because if the latency to reach DDR4 on this part looks like it did to go through the another die for Threadripper, then it's really not going to be that good of a part.
At this point, I think that we can say that the supposed leak passed around of a bunch of 7 nm Ryzen processors being announced at CES was completely fake.
At this point, I think that we can say that the supposed leak passed around of a bunch of 7 nm Ryzen processors being announced at CES was completely fake.
The timing of it yeah, i still think the actual details of the leak are correct, except the price, and the exact clocks, which will be finalized once they finish binning the chips. Oh and the leaks accurately predicted more than 8 cores on top of the chiplets as well.
Also AMD confirmed that the ES sample in the 9900k test was running up to 4.6Ghz@unconfirmed~75W, compared to ~125w for the 9900k, so that's 3 for 3 for the leaks atm.
Now the question is what AMD did to solve the latency problem of needing multiple hops to reach system memory. Or whether they solved it. Because if the latency to reach DDR4 on this part looks like it did to go through the another die for Threadripper, then it's really not going to be that good of a part.
This chiplet design is actually supposed to solve the latency problem, that's the main reason for it's existence. They way i understood it is compared to the possibility of multiple hops through CCX-es producing latency penalty on the old design, with the new chiplet design you are guaranteed to only do 1 additional hop before the RAM, and that's the hop to the I/O die, and that's it, no cross hoping like before. So at least 8 cores are latency free , i don't know what the impact will be when they add the second 8 core die.
And also there's rumors swirling that RAM will not impact the Infinity Fabric as much now, and that AMD is switching to a highly clocked uncore with ryzen 3.
Now the question is what AMD did to solve the latency problem of needing multiple hops to reach system memory. Or whether they solved it. Because if the latency to reach DDR4 on this part looks like it did to go through the another die for Threadripper, then it's really not going to be that good of a part.
This chiplet design is actually supposed to solve the latency problem, that's the main reason for it's existence. They way i understood it is compared to the possibility of multiple hops through CCX-es producing latency penalty on the old design, with the new chiplet design you are guaranteed to only do 1 additional hop before the RAM, and that's the hop to the I/O die, and that's it, no cross hoping like before. So at least 8 cores are latency free , i don't know what the impact will be when they add the second 8 core die.
And also there's rumors swirling that RAM will not impact the Infinity Fabric as much now, and that AMD is switching to a highly clocked uncore with ryzen 3.
Nonsense. Yields and production cost are the reason for the chiplet design, not latency. Moving from one chip to another physical chip is always going to be slower than moving the same distance within a chip. The only question is how much slower, how much more power it will consume, and how much worse it will be by any other metric you can think of.
If you want to build a 64-core server CPU with 8 DDR4 channels and 128 PCI Express 4.0 lanes on a cutting edge process node by making an enormous, monolithic die, then your yields are going to be terrible. Break that into a bunch of small chips and when something is defective, you can throw that one chiplet into the garbage without having to throw away the entire, enormous die.
Hopefully AMD has found a way to make it so that the chiplet approach adds not very much latency and not very much power. Adding 5 ns to your global memory latency is not the same problem that adding 50 ns would be. If they're going chiplets everywhere, then optimizing for this was surely a major focus of the design. But there's only so much you can do before physics stops you, and I'm not really sure what is possible here.
Hell, my I7 4790k keeps up with most of the "modern" cpus and it's what, 5 "generations" old now?
I'm rocking an overclocked 3570k Paired with a 1060 Can play all of today's games... maybe not with all the bells and whistles but everything runs fine...
All time classic MY NEW FAVORITE POST! (Keep laying those bricks)
"I should point out that no other company has shipped out a beta on a disc before this." - Official Mortal Online Lead Community Moderator
Proudly wearing the Harbinger badge since Dec 23, 2017.
Coined the phrase "Role-Playing a Development Team" January 2018
"Oddly Slap is the main reason I stay in these forums." - Mystichaze April 9th 2018
Hell, my I7 4790k keeps up with most of the "modern" cpus and it's what, 5 "generations" old now?
I'm rocking an overclocked 3570k Paired with a 1060 Can play all of today's games... maybe not with all the bells and whistles but everything runs fine...
the rig: i7/4790, GTX970, 16GB RAM. I have an ASROCK Z97 but my CPU isn't unlocked.
My biggest annoyance is the memory limit of the 970. Everything else feels pretty even. I want to upgrade my graphics card, but I don't need to upgrade anything at all. My system is also really stable and I like that. When upgrading that matters as much to me as the performance bump does. So far I have no reason to upgrade.
I'm pretty sure my next base system will be AMD because I'm at the place where locking CPU clocking behind SKUs pisses me right the hell off.
Hardware landscape is becoming weird to me because I'm also watching what they try and offer for a graphics card later this year or maybe next. Intel graphics traditionally work well on Linux and BSD and a more powerful dedicated desktop graphics card with open drivers could be huge. On the other hand the card offering could be a joke, the drivers could be closed, and it might only support Intel boards. Who knows.
Almost the same situation here - 4790k but not heavily overclocked, 980, 32G RAM.
Everything I play runs well - I have more issues with games supporting interface scaling at 4k that I do FPS at 4k. The computer I have is overkill for 90% of what I do with it. For the other 10% - it's still good enough to not need an upgrade.
I'd like to upgrade, but I don't need to upgrade. I'm thinking my next big purchase will be some nice monitors, and then build a new rig around that. Threadripper looks like it would be a really fun watercooled build to get around, but I don't want to throw that kind of money at a PC again.
I built a really nice 2700x system, but my son needed it more so I am back to my old 4670K running at 4ghz system with a 960 and most games still run fine. I only have 8gb of ddr3 memory, but still intend to build another AMD system so won't bother to waste money adding to that. I think I will wait until AMD's 3000 series debuts to build my next one. I hope to avoid Intel and Nvidia in my next build if possible,
I fall into - "buy best gaming CPU" dont' care if it's Intel/AMD
There are definitely people who stick to one brand.
As far as "intel" people - I guess there are many of those who only stick to Intel as that's all they've ever used - there are simply way more people like this, than AMD people who only use AMD cpu - especially in the desktop market.
Intel's desktop market share is over 85% - AMD has their work cut out for them in that segment.
Yep, Intel has lots of contracts that tie up PC makers with their hardware. Was looking at laptops the other day and could only find two with the AMD 2400G chips. Most of the others had Intel cpus with the HD 630 graphics which sucks at about anything that requires graphics.
I fall into - "buy best gaming CPU" dont' care if it's Intel/AMD
There are definitely people who stick to one brand.
As far as "intel" people - I guess there are many of those who only stick to Intel as that's all they've ever used - there are simply way more people like this, than AMD people who only use AMD cpu - especially in the desktop market.
Intel's desktop market share is over 85% - AMD has their work cut out for them in that segment.
Yep, Intel has lots of contracts that tie up PC makers with their hardware. Was looking at laptops the other day and could only find two with the AMD 2400G chips. Most of the others had Intel cpus with the HD 630 graphics which sucks at about anything that requires graphics.
AMD has two big problems in laptops:
1) Their chips use more power at idle than Intel's, which hurts battery life. That's makes them a non-starter for a lot of purposes, and at minimum, a big problem if you weren't going to use their integrated GPU heavily. 2) AMD doesn't release drivers for their laptop GPUs. That's a showstopper if you were hoping to make heavy use out of that integrated GPU.
AMD has promised to fix the latter problem sometime in Q1 of this year. It's not clear when they'll address the former. Neither problem is relevant to desktops, however.
Let's put that into context. Ever since the launch of Ryzen, Cinebench has been a very favorable benchmark to AMD. It existed and was commonly used in CPUs well before Ryzen launched, so it's not like it's a benchmark created by AMD, but it is more favorable to AMD than most other benchmarks.
Cinebench has two versions: single-threaded and multi-threaded. The latter scales well to many CPU cores. On Anandtech, a Core i7-9900K beats a Ryzen 7 2700X by 23% at the single-threaded version and 26% at the multi-threaded one. The Intel CPU also uses a lot more power than the AMD one, so it's not like this performance comes for free. But for Zen 2 to win outright at the latter is a major improvement over the Ryzen 7 2700X.
It's nearly guaranteed that Zen 2 is going to be a lot more energy efficient at load than Coffee Lake, largely because of the process node. So while it's likely that AMD will still have a modest single-threaded performance deficit, it's also likely that AMD is able to keep the eight-core turbo clocks closer to the single-core turbo than Intel is able to when pushing all those cores precisely because it uses so much less power. That's probably how AMD wins at Cinebench.
But there are two other things that are critical to remember. One is that Intel charges a lot for that Core i7-9900K. If you can't afford to spend over $500 on a CPU and buy something that tops out at 4.0 GHz rather than the 5.0 GHz of the Core i7-9900K, Intel's single-threaded performance advantage nearly vanishes. Most people shop for the best CPU (or GPU or whatever) in a particular price range, not the best without regard to price.
The other issue is the Zen 2 dies that have been announced are primarily a server product, not a desktop one. AMD isn't ignoring the desktop market, but it's not their primary focus here. That reduced power consumption (due largely to the process node) means that AMD can pack in a lot more CPU cores for server workloads that scale well to many cores. That's where AMD is going to below Intel out of the water. Intel will still be competitive on desktops after Zen 2 launches, at least if you don't care much about power consumption.
I fall into - "buy best gaming CPU" dont' care if it's Intel/AMD
There are definitely people who stick to one brand.
As far as "intel" people - I guess there are many of those who only stick to Intel as that's all they've ever used - there are simply way more people like this, than AMD people who only use AMD cpu - especially in the desktop market.
Intel's desktop market share is over 85% - AMD has their work cut out for them in that segment.
Broadly agree, however, over the years I have found that the total thermal output from an AMD based system is consistently, and significantly, higher than Intel based systems. And with summer temperatures in the high 30's ( over 100 Fahrenheit) this is an important consideration. Run a couple of AMD boxes in a small room and you are going to want a dedicated room air conditioner.
I fall into - "buy best gaming CPU" dont' care if it's Intel/AMD
There are definitely people who stick to one brand.
As far as "intel" people - I guess there are many of those who only stick to Intel as that's all they've ever used - there are simply way more people like this, than AMD people who only use AMD cpu - especially in the desktop market.
Intel's desktop market share is over 85% - AMD has their work cut out for them in that segment.
Broadly agree, however, over the years I have found that the total thermal output from an AMD based system is consistently, and significantly, higher than Intel based systems. And with summer temperatures in the high 30's ( over 100 Fahrenheit) this is an important consideration. Run a couple of AMD boxes in a small room and you are going to want a dedicated room air conditioner.
Being a process node behind typically means that every time a transistor does something, it uses about 40% more power than if you weren't behind. For decades, AMD has been behind on process nodes. This year, if Zen 2 launches around the middle of this year as AMD says it will, AMD is going to be ahead on process nodes for the first time ever. And that means that if your goal is to keep load power consumption down at a given level of performance, AMD will be your only option for the CPU. Possibly also the GPU, if Navi is good, at least until Nvidia gets to 7 nm. Rumors say that Nvidia won't get there until 2020.
I fall into - "buy best gaming CPU" dont' care if it's Intel/AMD
There are definitely people who stick to one brand.
As far as "intel" people - I guess there are many of those who only stick to Intel as that's all they've ever used - there are simply way more people like this, than AMD people who only use AMD cpu - especially in the desktop market.
Intel's desktop market share is over 85% - AMD has their work cut out for them in that segment.
Broadly agree, however, over the years I have found that the total thermal output from an AMD based system is consistently, and significantly, higher than Intel based systems. And with summer temperatures in the high 30's ( over 100 Fahrenheit) this is an important consideration. Run a couple of AMD boxes in a small room and you are going to want a dedicated room air conditioner.
Being a process node behind typically means that every time a transistor does something, it uses about 40% more power than if you weren't behind. For decades, AMD has been behind on process nodes. This year, if Zen 2 launches around the middle of this year as AMD says it will, AMD is going to be ahead on process nodes for the first time ever. And that means that if your goal is to keep load power consumption down at a given level of performance, AMD will be your only option for the CPU. Possibly also the GPU, if Navi is good, at least until Nvidia gets to 7 nm. Rumors say that Nvidia won't get there until 2020.
I said total thermal output and that is what I meant, for a given amount of "work" AMD systems have a significantly higher total thermal output.
Being a process node behind typically means that every time a transistor does something, it uses about 40% more power than if you weren't behind. For decades, AMD has been behind on process nodes. This year, if Zen 2 launches around the middle of this year as AMD says it will, AMD is going to be ahead on process nodes for the first time ever. And that means that if your goal is to keep load power consumption down at a given level of performance, AMD will be your only option for the CPU. Possibly also the GPU, if Navi is good, at least until Nvidia gets to 7 nm. Rumors say that Nvidia won't get there until 2020.
I said total thermal output and that is what I meant, for a given amount of "work" AMD systems have a significantly higher total thermal output.
The laws of physics don't care what logo is on the box that a CPU came in. It turns out that they do care tremendously about things like voltage, current, capacitance, and resistance. What you likely mean is that doing a given amount of work uses more power on an inferior process node than it would on a better one. For decades, AMD has been stuck on process nodes that were inferior to Intel's. That, and not the logo on the box, is probably what led to your observation. This year, that's going to flip around and AMD will be on a superior process node to Intel for the first time ever.
The process node isn't the only thing that affects power output, of course. Some architectures are simply better than others, and AMD's Bulldozer architecture, or going back further, Intel's NetBurst (Pentium 4 and Pentium D) were notorious power hogs. But right now, AMD and Intel are about even on architectural goodness. It's certainly close enough that if either side were a full process node ahead, they'd win by a lot on efficiency.
Unless the issue is that you don't realize that thermal output is the same as power consumption. That's just conservation of energy.
Or if you're talking specifically about chipsets as opposed to CPUs, then you have a point, but probably don't realize what it is.
I fall into - "buy best gaming CPU" dont' care if it's Intel/AMD
There are definitely people who stick to one brand.
As far as "intel" people - I guess there are many of those who only stick to Intel as that's all they've ever used - there are simply way more people like this, than AMD people who only use AMD cpu - especially in the desktop market.
Intel's desktop market share is over 85% - AMD has their work cut out for them in that segment.
Broadly agree, however, over the years I have found that the total thermal output from an AMD based system is consistently, and significantly, higher than Intel based systems. And with summer temperatures in the high 30's ( over 100 Fahrenheit) this is an important consideration. Run a couple of AMD boxes in a small room and you are going to want a dedicated room air conditioner.
Being a process node behind typically means that every time a transistor does something, it uses about 40% more power than if you weren't behind. For decades, AMD has been behind on process nodes. This year, if Zen 2 launches around the middle of this year as AMD says it will, AMD is going to be ahead on process nodes for the first time ever. And that means that if your goal is to keep load power consumption down at a given level of performance, AMD will be your only option for the CPU. Possibly also the GPU, if Navi is good, at least until Nvidia gets to 7 nm. Rumors say that Nvidia won't get there until 2020.
I said total thermal output and that is what I meant, for a given amount of "work" AMD systems have a significantly higher total thermal output.
Not true at all. I have had multiple AMD systems running in my apartment and noticed no increase in temperature. The new Zen 2 cpus according to what we know we run even cooler.
Interesting test by HardOCP shows a 7700k CPU outperforming a 9900k CPU in Battlefield V. It also shows how much ray tracing degrades performance even with a 2080ti. Makes a good point these games are far more GPU limited than CPU.
Interesting test by HardOCP shows a 7700k CPU outperforming a 9900k CPU in Battlefield V. It also shows how much ray tracing degrades performance even with a 2080ti. Makes a good point these games are far more GPU limited than CPU.
Even in 1080p, ray tracing is a bit too demanding at this point in time. Can't imagine how bad ray tracing would be with a 2060.
So for a gamer, the 9900K seems like a big waste of money.
7700K was overclocked. I don't think it's that surprising, rather it shows that single core performance is still important, and both Intel and AMD pack their best processors with more cores than most games can take advantage of.
The difference between the 6000, 7000, 800 and 9000 series Intel processors is ... the numbers 6, 7, 8 and 9. Those are all Skylake-based, and the differences are minimal generation over generation. Hence Quiz's title for the OP - Skylake Refresh Refresh Refresh.
So you would completely expect that, between any of those CPUs, that whichever one has the clock speed advantage would come up the winner in a core-limited race. And that's what HardOCP showed - the 7700 was OCed faster than the 9900 stock, and the benchmarks tell the story.
nVidia's (maybe not the company itself but at least the white knights) were claiming that Raytracing was supposed to be run on a 9000 series CPU and the testing HardOCP did with a 7000 series was invalid because of that. The HardOCP team also knows these are all Skylake CPUs... and they did the testing to prove that ... no, that wasn't what was limiting Raytracing performance, and that for a given architecture clock speed is going to win out if core count isn't the deciding factor.
Being a process node behind typically means that every time a transistor does something, it uses about 40% more power than if you weren't behind. For decades, AMD has been behind on process nodes. This year, if Zen 2 launches around the middle of this year as AMD says it will, AMD is going to be ahead on process nodes for the first time ever. And that means that if your goal is to keep load power consumption down at a given level of performance, AMD will be your only option for the CPU. Possibly also the GPU, if Navi is good, at least until Nvidia gets to 7 nm. Rumors say that Nvidia won't get there until 2020.
I said total thermal output and that is what I meant, for a given amount of "work" AMD systems have a significantly higher total thermal output.
The laws of physics don't care what logo is on the box that a CPU came in. It turns out that they do care tremendously about things like voltage, current, capacitance, and resistance. What you likely mean is that doing a given amount of work uses more power on an inferior process node than it would on a better one. For decades, AMD has been stuck on process nodes that were inferior to Intel's. That, and not the logo on the box, is probably what led to your observation. This year, that's going to flip around and AMD will be on a superior process node to Intel for the first time ever.
The process node isn't the only thing that affects power output, of course. Some architectures are simply better than others, and AMD's Bulldozer architecture, or going back further, Intel's NetBurst (Pentium 4 and Pentium D) were notorious power hogs. But right now, AMD and Intel are about even on architectural goodness. It's certainly close enough that if either side were a full process node ahead, they'd win by a lot on efficiency.
Unless the issue is that you don't realize that thermal output is the same as power consumption. That's just conservation of energy.
Or if you're talking specifically about chipsets as opposed to CPUs, then you have a point, but probably don't realize what it is.
Thanks for the condescension. Yes I am entirely aware of the physics, probably better than you, having worked in hardware development. But my point is that while both brands are capable of excellent performance, historically systems built with AMD CPU's have a higher measurable thermal output. And this has an impact on room temperature. Glad to hear that AMDs latest generation may reverse that.
Comments
https://www.techradar.com/news/amd-is-now-selling-twice-as-many-processors-as-intel
Most of the people I build PCs for these days don't want Intel and prefer AMD, but if you need top end single thread power, Intel is still the leader.
So much for non-guaranteed chiplet design, and only threadripper chiplet design xD
Now the question is what AMD did to solve the latency problem of needing multiple hops to reach system memory. Or whether they solved it. Because if the latency to reach DDR4 on this part looks like it did to go through the another die for Threadripper, then it's really not going to be that good of a part.
At this point, I think that we can say that the supposed leak passed around of a bunch of 7 nm Ryzen processors being announced at CES was completely fake.
Oh and the leaks accurately predicted more than 8 cores on top of the chiplets as well.
Also AMD confirmed that the ES sample in the 9900k test was running up to 4.6Ghz@unconfirmed~75W, compared to ~125w for the 9900k, so that's 3 for 3 for the leaks atm.
"My Fantasy is having two men at once...
One Cooking and One Cleaning!"
---------------------------
"A good man can make you feel sexy,
strong and able to take on the whole world...
oh sorry...that's wine...wine does that..."
They way i understood it is compared to the possibility of multiple hops through CCX-es producing latency penalty on the old design, with the new chiplet design you are guaranteed to only do 1 additional hop before the RAM, and that's the hop to the I/O die, and that's it, no cross hoping like before. So at least 8 cores are latency free
And also there's rumors swirling that RAM will not impact the Infinity Fabric as much now, and that AMD is switching to a highly clocked uncore with ryzen 3.
If you want to build a 64-core server CPU with 8 DDR4 channels and 128 PCI Express 4.0 lanes on a cutting edge process node by making an enormous, monolithic die, then your yields are going to be terrible. Break that into a bunch of small chips and when something is defective, you can throw that one chiplet into the garbage without having to throw away the entire, enormous die.
Hopefully AMD has found a way to make it so that the chiplet approach adds not very much latency and not very much power. Adding 5 ns to your global memory latency is not the same problem that adding 50 ns would be. If they're going chiplets everywhere, then optimizing for this was surely a major focus of the design. But there's only so much you can do before physics stops you, and I'm not really sure what is possible here.
Paired with a 1060
Can play all of today's games... maybe not with all the bells and whistles but everything runs fine...
All time classic MY NEW FAVORITE POST! (Keep laying those bricks)
"I should point out that no other company has shipped out a beta on a disc before this." - Official Mortal Online Lead Community Moderator
Proudly wearing the Harbinger badge since Dec 23, 2017.
Coined the phrase "Role-Playing a Development Team" January 2018
"Oddly Slap is the main reason I stay in these forums." - Mystichaze April 9th 2018
Everything I play runs well - I have more issues with games supporting interface scaling at 4k that I do FPS at 4k. The computer I have is overkill for 90% of what I do with it. For the other 10% - it's still good enough to not need an upgrade.
I'd like to upgrade, but I don't need to upgrade. I'm thinking my next big purchase will be some nice monitors, and then build a new rig around that. Threadripper looks like it would be a really fun watercooled build to get around, but I don't want to throw that kind of money at a PC again.
https://www.hardocp.com/news/2019/01/11/amd_cinebench_benchmark_demo_at_ces_2019_buries_current_intel_lineup/
Read it and weep for you Intel people.
1) Their chips use more power at idle than Intel's, which hurts battery life. That's makes them a non-starter for a lot of purposes, and at minimum, a big problem if you weren't going to use their integrated GPU heavily.
2) AMD doesn't release drivers for their laptop GPUs. That's a showstopper if you were hoping to make heavy use out of that integrated GPU.
AMD has promised to fix the latter problem sometime in Q1 of this year. It's not clear when they'll address the former. Neither problem is relevant to desktops, however.
If you are holding out for the perfect game, the only game you play will be the waiting one.
Cinebench has two versions: single-threaded and multi-threaded. The latter scales well to many CPU cores. On Anandtech, a Core i7-9900K beats a Ryzen 7 2700X by 23% at the single-threaded version and 26% at the multi-threaded one. The Intel CPU also uses a lot more power than the AMD one, so it's not like this performance comes for free. But for Zen 2 to win outright at the latter is a major improvement over the Ryzen 7 2700X.
It's nearly guaranteed that Zen 2 is going to be a lot more energy efficient at load than Coffee Lake, largely because of the process node. So while it's likely that AMD will still have a modest single-threaded performance deficit, it's also likely that AMD is able to keep the eight-core turbo clocks closer to the single-core turbo than Intel is able to when pushing all those cores precisely because it uses so much less power. That's probably how AMD wins at Cinebench.
But there are two other things that are critical to remember. One is that Intel charges a lot for that Core i7-9900K. If you can't afford to spend over $500 on a CPU and buy something that tops out at 4.0 GHz rather than the 5.0 GHz of the Core i7-9900K, Intel's single-threaded performance advantage nearly vanishes. Most people shop for the best CPU (or GPU or whatever) in a particular price range, not the best without regard to price.
The other issue is the Zen 2 dies that have been announced are primarily a server product, not a desktop one. AMD isn't ignoring the desktop market, but it's not their primary focus here. That reduced power consumption (due largely to the process node) means that AMD can pack in a lot more CPU cores for server workloads that scale well to many cores. That's where AMD is going to below Intel out of the water. Intel will still be competitive on desktops after Zen 2 launches, at least if you don't care much about power consumption.
The process node isn't the only thing that affects power output, of course. Some architectures are simply better than others, and AMD's Bulldozer architecture, or going back further, Intel's NetBurst (Pentium 4 and Pentium D) were notorious power hogs. But right now, AMD and Intel are about even on architectural goodness. It's certainly close enough that if either side were a full process node ahead, they'd win by a lot on efficiency.
Unless the issue is that you don't realize that thermal output is the same as power consumption. That's just conservation of energy.
Or if you're talking specifically about chipsets as opposed to CPUs, then you have a point, but probably don't realize what it is.
https://www.hardocp.com/article/2019/01/13/battlefield_v_nvidia_ray_tracing_i99900k_cpu_testing/1
Even in 1080p, ray tracing is a bit too demanding at this point in time. Can't imagine how bad ray tracing would be with a 2060.
So for a gamer, the 9900K seems like a big waste of money.
So you would completely expect that, between any of those CPUs, that whichever one has the clock speed advantage would come up the winner in a core-limited race. And that's what HardOCP showed - the 7700 was OCed faster than the 9900 stock, and the benchmarks tell the story.
nVidia's (maybe not the company itself but at least the white knights) were claiming that Raytracing was supposed to be run on a 9000 series CPU and the testing HardOCP did with a 7000 series was invalid because of that. The HardOCP team also knows these are all Skylake CPUs... and they did the testing to prove that ... no, that wasn't what was limiting Raytracing performance, and that for a given architecture clock speed is going to win out if core count isn't the deciding factor.
거북이는 목을 내밀 때 안 움직입니다