Image: AMD

During last week’s Zen 3 event, AMD suggested that its flagship Ryzen 5000 Series processor would be eating Intel’s for breakfast in both gaming and content creation. While those claims were strictly in-house (i.e., subject to shenanigans), we now have third-party metrics that suggest red team’s performance data is on point.

The first Cinebench R20 benchmarks for the Ryzen 9 5950X can be found on CPU Monkey (via Guru 3D), and they are definitely above and beyond what Intel’s flagship can muster. In the single-core test, AMD’s Ryzen 9 5950X scored 604, while the Core i9-10900KF only managed 539 (11 percent difference).

The multi-core test was also (unsurprisingly) in AMD’s favor, as its 16C/32T beast managed to achieve a score of 10,360. In comparison, Intel’s 10C/20T Core i9-10900KF only hit 6,399 (47 percent difference).

You can check out the rest of the comparisons below. AMD’s Ryzen 9 5950X (and the rest of the Ryzen 5000 Series lineup) will be released on November 5.

CPUCores / threadsBase / boost (GHz)R20 Single-CoreR20 Multi-Core
AMD Ryzen 9 5950X16/323.40 / 4.9064110,360
AMD Ryzen 9 5900X12/243.70 / 4.806298,168
AMD Ryzen 7 5800X8/163.80 / 4.706185,724
AMD Ryzen 5 5600X6/123.70 / 4.606044,312
Intel Core i7-1185G74/83.00 / 4.805982,477
Intel Core i7-1165G74/82.80 / 4.705612,234
AMD Ryzen 7 3800XT8/164.20 / 4.705395,122
AMD Ryzen 9 3900XT12/244.10 / 4.705397,244
Intel Core i9-10900KF10/203.70 / 5.305396,399
Intel Core i9-10900K10/203.70 / 5.305396,399

Recent Posts

Join the Conversation

35 Comments

  1. I’m interested in the 5900x, but will have to change mobo’s.

    Maybe after the first of the year

  2. More ignorance from the industry & consumers! It is not fair or objective to compare a 16 core cpu with a 10 core and expect similar results. AMD fanboys don’t get this because they allow their personal feelings to dedicate their opinions. So of course they believe AMD is winning when garbage articles like this one show up.

    To be objective, compare an 8 core Intel cpu with an 8 core AMD such as a 9900k/10700k with a 3800x or 5800x! Sorry so many AMD fanboys can’t understand why this is the more objective thing to do.

    Since Intel doesn’t have a 12 or 16 core cpu for their mainstream platform, compare the 10 core cpus. BUT of course AMD 12 & 16 core cpus will blow away Intel’s 10 core, because if they didn’t, it would look super bad for AMD!

    1. Funny comment. I as a customer am interested only in performance per invested dolar. So if we compare 500 euros 10900kf vs 550 euro 5900 i get like 30 % more of performance for like 10% more money. If take this further if i wanna get intel to performe as 5950 i have to buy two intel’s cpus two mobs two sets of RAM two sets of gpus two psus two cases and probably the most expensive two licenses of archicad two licenses of 3ds max two licenses of vray etc. If we take those expanses to calculation you will see that AMDs 5950 is a bargain and is beating Intel.
      I dont care how many cores are working for me when job is done on time. Core vs core is only interesting to cpu engineers and those fanboys like you.
  3. More ignorance from the industry & consumers! It is not fair or objective to compare a 16 core cpu with a 10 core and expect similar results. AMD fanboys don’t get this because they allow their personal feelings to dedicate their opinions. So of course they believe AMD is winning when garbage articles like this one show up.

    To be objective, compare an 8 core Intel cpu with an 8 core AMD such as a 9900k/10700k with a 3800x or 5800x! Sorry so many AMD fanboys can’t understand why this is the more objective thing to do.

    Since Intel doesn’t have a 12 or 16 core cpu for their mainstream platform, compare the 10 core cpus. BUT of course AMD 12 & 16 core cpus will blow away Intel’s 10 core, because if they didn’t, it would look super bad for AMD!

    Did you just blatantly ignore the single core scores?

  4. That 5950x single core score is beastly.

    I hope it is able to maintain it better than my 3930x can. It doesn’t take much to knock the 3930k off of max boost mode, including almost any background task, which turns it into a little bit of an academic number, rather than a real use one.

    Also, why is it only the insane 32 core models come with the best boost clocks?

    Why won’t they learn that huge core count and high clocks generally have different applications?

    I’d like an 8 core model with the max boost clock of the 32 core model. Who the hell needs all those added 24 cores?

  5. That 5950x single core score is beastly.

    I hope it is able to maintain it better than my 3930x can. It doesn’t take much to knock the 3930k off of max boost mode, including almost any background task, which turns it into a little bit of an academic number, rather than a real use one.

    Also, why is it only the insane 32 core models come with the best boost clocks?

    Why won’t they learn that huge core count and high clocks generally have different applications?

    I’d like an 8 core model with the max boost clock of the 32 core model. Who the hell needs all those added 24 cores?

    Well theyre establishing price points, you want high clocks, pay up :)

  6. Well theyre establishing price points, you want high clocks, pay up :)

    Yeah, but it’s sort of like telling a sports car enthusiast, that if he wants high horsepower, he’s going to have to buy a dump truck.

    They are different products for different users.

  7. More ignorance from the industry & consumers! It is not fair or objective to compare a 16 core cpu with a 10 core and expect similar results. AMD fanboys don’t get this because they allow their personal feelings to dedicate their opinions. So of course they believe AMD is winning when garbage articles like this one show up.

    To be objective, compare an 8 core Intel cpu with an 8 core AMD such as a 9900k/10700k with a 3800x or 5800x! Sorry so many AMD fanboys can’t understand why this is the more objective thing to do.

    Since Intel doesn’t have a 12 or 16 core cpu for their mainstream platform, compare the 10 core cpus. BUT of course AMD 12 & 16 core cpus will blow away Intel’s 10 core, because if they didn’t, it would look super bad for AMD!

    I would think comparing by price point would be fair.

  8. More ignorance from the industry & consumers! It is not fair or objective to compare a 16 core cpu with a 10 core and expect similar results. AMD fanboys don’t get this because they allow their personal feelings to dedicate their opinions. So of course they believe AMD is winning when garbage articles like this one show up.

    To be objective, compare an 8 core Intel cpu with an 8 core AMD such as a 9900k/10700k with a 3800x or 5800x! Sorry so many AMD fanboys can’t understand why this is the more objective thing to do.

    Since Intel doesn’t have a 12 or 16 core cpu for their mainstream platform, compare the 10 core cpus. BUT of course AMD 12 & 16 core cpus will blow away Intel’s 10 core, because if they didn’t, it would look super bad for AMD!

    Looking at the comparisons posted, I would agree that they weren’t the best choices. Finding a comparison for reviews can be challenging – we tend to go towards price as the arbiter of comparison, so in this case, the 10900 is the comparison to the 5900x as they are in the same ballpark.

    For the 5950x, that’s more difficult – you have to move to the Intel workstation platform to get more cores, but there’s also platform cost differences since you’re getting out of the consumer space. There really is no good comparison there other than maybe the 3950x.

    I suppose the single core tests will be fine for now… Meanwhile we wait for reviews!

  9. Yeah, but it’s sort of like telling a sports car enthusiast, that if he wants high horsepower, he’s going to have to buy a dump truck.

    They are different products for different users.

    Nah, he just needs to buy the "S" model of said sporstcar….(ask Porsche)

  10. Did nobody notice he completely has the wrong numbers in this story? Look at the chart, it’s the 5600x that got 604, not the 5950x… that scored 641. At least he got the right MT score I guess ;).
  11. Did nobody notice he completely has the wrong numbers in this story? Look at the chart, it’s the 5600x that got 604, not the 5950x… that scored 641. At least he got the right MT score I guess ;).

    Yep, according to the chart you are correct.

  12. More ignorance from the industry & consumers! It is not fair or objective to compare a 16 core cpu with a 10 core and expect similar results. AMD fanboys don’t get this because they allow their personal feelings to dedicate their opinions. So of course they believe AMD is winning when garbage articles like this one show up.

    To be objective, compare an 8 core Intel cpu with an 8 core AMD such as a 9900k/10700k with a 3800x or 5800x! Sorry so many AMD fanboys can’t understand why this is the more objective thing to do.

    Since Intel doesn’t have a 12 or 16 core cpu for their mainstream platform, compare the 10 core cpus. BUT of course AMD 12 & 16 core cpus will blow away Intel’s 10 core, because if they didn’t, it would look super bad for AMD!

    lol, this is rich. It’s not fair to compare top end to top end, or even 2 items in the same price range. Let me guess, you were against comparing AMD’s 5700xt to anything above $400 too? Yah, that’s what I thought, if you want to stay in your vacuum that’s fine, the rest of us live in the real world. Amazing how quick Intel sends the shills out before benchmarks even start showing up. With a name like Amdguru, you aren’t fooling anyone.

  13. I would think comparing by price point would be fair.

    It’s really the only comparison that matters.

    If I spend X dollars, what can I get for it, and how will it perform.

    Other comparisons can be interesting from an academic perspective to determine all sorts of things like performance per core, per clock, etc. but in th end, where the rubber meets the road, all that matters is, how did it perform for the money.

  14. I might actually wait around for the fire sale on Zen 2 parts, and pick up a 3800x or 3800xt as a last hurrah upgrade for my stepsons B350 Tomahawk motherboard.

    I wonder what the Zen3 Threadrippers would bring.

    I opted for a Threadripper my last upgrade because the regular Ryzens didn’t have enough PCIe lanes for me, not because I needed or wanted the extra cores.

    If I had my druthers I’d be able to buy a Zen3 Threadripper with all the PCIe lanes and same max clocks as the 5950x, but only 8C16T. I know this will never happen, but this would essentially be the ideal CPU for me. I have no need for all these crazy core counts, but I still like my expansion.

    I hate how they have to tie high PCIe lanes, high core counts and high clock speeds together. It’s really frustrating.

  15. More ignorance from the industry & consumers! It is not fair or objective to compare a 16 core cpu with a 10 core and expect similar results. AMD fanboys don’t get this because they allow their personal feelings to dedicate their opinions. So of course they believe AMD is winning when garbage articles like this one show up.

    To be objective, compare an 8 core Intel cpu with an 8 core AMD such as a 9900k/10700k with a 3800x or 5800x! Sorry so many AMD fanboys can’t understand why this is the more objective thing to do.

    Since Intel doesn’t have a 12 or 16 core cpu for their mainstream platform, compare the 10 core cpus. BUT of course AMD 12 & 16 core cpus will blow away Intel’s 10 core, because if they didn’t, it would look super bad for AMD!

    I’m an AMD fanboy … but I don’t quite fit into your little box there .. As Zarathustra mentioned .. what does my hard earned dollar get me that gives ME the best bang for MY buck .. right now, it looks like AMD’s offerings and even more so their future offerings are what gives ME what I want .. :)

    1. I get your response. My issue lies w/ the fact that, as a gamer, core counts that high are pointless. The fact that their best single core performance is locked behind their most expensive cpu is a bit baffling. You can argue for a 6 core or 8 core, even a 10 core as future proofing. A 16 core cpu though? That’s overkill even for streaming and rendering for most people.

      Its also going to come down to a matter of head room for oc’ing. They said these cpus will have more head room, but more compared to what? Its been shown that oc’ing zen 2 can offer very, very little or no improvement. The built in boost, w/e its called, tends to outperform even manual oc’s.

      Sadly, there’s really no sure-fire gaming benchmark. In gaming, Intel cpus tend to marginally edge out AMD for some reason when all things are as equal as they can be. Some times even when AMD should have a clear edge. PCIe 4.0 really hasn’t made itself necessary yet as really only ssds ‘need’ it atm. Though the real world application of even those ssds will only save you a few minutes over the course of a year.

      In the end, cpu’s are so far ahead of gpus that you have to be pushing insane framerates at lower resolutions to become cpu bound. At which point you should likely be looking to upgrade your display and using a higher resolution. As there is a legit argument that those actually playing at such high framerates do so more out of ‘superstition’ than any actual gaming performance. Much like athletes and their lucky X, refusing to shave on winning streaks, etc.

      1. Funny comment. I as a customer am interested only in performance per invested dolar. So if we compare 500 euros 10900kf vs 550 euro 5900 i get like 30 % more of performance for like 10% more money. If take this further if i wanna get intel to performe as 5950 i have to buy two intel’s cpus two mobs two sets of RAM two sets of gpus two psus two cases and probably the most expensive two licenses of archicad two licenses of 3ds max two licenses of vray etc. If we take those expanses to calculation you will see that AMDs 5950 is a bargain and is beating Intel.
        I dont care how many cores are working for me when job is done on time. Core vs core is only interesting to cpu engineers and those fanboys like you.
  16. Single Core = beautiful
    Multi Core = beautiful
    Price to performance is the only thing that really matters, you don’t match up core count without first checking price to performance. Can’t wait to see third party reviews and benchmarks.
  17. This isn’t surprising. If you look at the benchmark data in the post, it shows AMD was already slightly ahead with the 3000 series so long as the base and boost clocks were high enough on the single thread test. In the multithread test, AMD has been cleaning house on that front since the Ryzen 9 3900X launched last year.

    The 8c/16t AMD 3800XT doesn’t match the 10c/20t Intel part, which isn’t surprising given the core deficit and lower price point. On the price front, AMD generally does better either by offering 12c/16t parts for less than Intel’s 10c/20t parts or by offering 8c/16t for less than Intel does and outperforming them much of the time. The 3900X, 3900XT and 3950X beat Intel’s best in the multi-thread test with ease. The 3950X, when pitted against the Core i9-10980XE does better than you’d think unless the latter is overclocked and then it’s game over. When pushed high enough, it beats the 3950X across the board in virtually every test.

    But as David said, comparing costs places the 3950X and 10980XE in different leagues. It’s the same when you move up the stack. The 10980XE is it for Intel and it can’t remotely match anything that AMD has above it. The 3960X, 3970X, 3990X etc. are all considerably faster. The baby of the third generation Threadripper line is about $400 more expensive than the 10980XE to start, but the platform costs are similar if not higher. Again, it places them in different leagues. The rest of the Threadripper line doesn’t even make sense to compare to the 10980XE as the price not only doubles (or quadruples it), but the core counts are considerably greater as well.

    You have the 3950X which is more expensive than anything Intel has in the mainstream segment but way less than Intel’s best HEDT offering. It trounces anything but the 10980XE from Intel’s HEDT product line. However, the 10980XE sort of follows suit by being way cheaper than anything AMD offers in the third generation Threadripper family. The 10980XE is probably still a little too expensive for what it is, but it sits between the 3950X and 3960X in both price and performance. Again, if price or core counts are your metric for comparison, than the 3950X and 10980XE really are islands of their own.

  18. Funny comment. I as a customer am interested only in performance per invested dolar. So if we compare 500 euros 10900kf vs 550 euro 5900 i get like 30 % more of performance for like 10% more money. If take this further if i wanna get intel to performe as 5950 i have to buy two intel’s cpus two mobs two sets of RAM two sets of gpus two psus two cases and probably the most expensive two licenses of archicad two licenses of 3ds max two licenses of vray etc. If we take those expanses to calculation you will see that AMDs 5950 is a bargain and is beating Intel.
    I dont care how many cores are working for me when job is done on time. Core vs core is only interesting to cpu engineers and those fanboys like you.

    As a customer I tend to only care about who has the highest performance. Price per dollar is secondary. For example: The 3960X is faster than the 3950X, but the $800 increase isn’t something I’d opt for given that it won’t benefit me in gaming and I can’t leverage the 3960X’s extra performance in the non-gaming applications I run, so it doesn’t make sense to pony up the extra for it. Similarly, the RTX Titan wasn’t worth it to me over the cost of the RTX 2080 Ti.

    Those are two examples where I didn’t opt to buy something simply because it was the top of the stack. However, I generally do just that. I’ve bought plenty of Extreme Edition CPU’s over the years. I almost always buy the fastest video card available in each generation. I’ve bought plenty of Titans and Ti cards in pairs for SLI. Of course, I justify that because I typically go for gaming resolutions far beyond 1080P, and every bit helps when you do that.

  19. As a customer I tend to only care about who has the highest performance. Price per dollar is secondary. For example: The 3960X is faster than the 3950X, but the $800 increase isn’t something I’d opt for given that it won’t benefit me in gaming and I can’t leverage the 3960X’s extra performance in the non-gaming applications I run, so it doesn’t make sense to pony up the extra for it. Similarly, the RTX Titan wasn’t worth it to me over the cost of the RTX 2080 Ti.

    Those are two examples where I didn’t opt to buy something simply because it was the top of the stack. However, I generally do just that. I’ve bought plenty of Extreme Edition CPU’s over the years. I almost always buy the fastest video card available in each generation. I’ve bought plenty of Titans and Ti cards in pairs for SLI. Of course, I justify that because I typically go for gaming resolutions far beyond 1080P, and every bit helps when you do that.

    I’d agree that things like FPS/Dollar or Benchmark Score/Dollar are generally useless. If this were the best metric, we’d all buy low end chips all the time, as price tends to increase exponentially with performance.

    I tend to see it as this:

    I have a vague picture of what I am OK with spending. That picture is slightly malleable, but not by much, at least not by a lot very quickly.

    Once I’ve established roughly what I am willing to pay, I’ll look for the fastest hardware in the workloads I care about that fits within that budget.

    It’s really quite simple.

    That said, I do use some of these other benchmarks to help predict performance, if the workloads I care about arent well benchmarked (and based on my tastes, they rarely are, and as we all know, real world benchmarks in the actual software being run are always superior.

    Provided a CPU has at least 6 to 8 cores, (I t used to be at least 4, then at least 4-6, at least 6, and now we are at at least 6-8, maybe on the border of moving towards at least 8. Not sure about that one yet.)

    Anyway, I digress. Back to where I was. Provided a CPU has at least 6 to 8 cores, I find that the best predictive benchmark is one that is lightly threaded, as most games are still fairly lightly threaded.

    Berfore the launch of Ryzen I used to be a huge proponent of the single threaded Cinebench benchmark as a good predictor, but Ryzen has turned that on its head, as – at least in my experience – Ryzen chips seem to be knocked out of their max boosts very easily by other things going on on the chip. So a single threaded bench tends to over-inflate things compared to reality, because in reality there is almost always something going on on another thread, that tends to knock it out of max boost.

    I’d like to see a bench like Cinebench, that loads up one core 100%, loads up the second core 50%, the third core 25%, the fourth core 12.5%, fifth core 12.25%, and so on until about 8 cores or so, and then loads the rest at 1-2%.

    I think that would be a pretty awesome (relative, not absolute) predictor of "the average game" performance.

  20. I’d agree that things like FPS/Dollar or Benchmark Score/Dollar are generally useless. If this were the best metric, we’d all by low end chips all the time, as price tends to increase exponentially with performance.

    I tend to see it as this:

    I have a vague picture of what I am OK with spending. That picture is slightly malleable, but not by much, at least not by a lot very quickly.

    Once I’ve established roughly what I am willing to pay, I’ll look for the fastest hardware in the workloads I care about that fits within that budget.

    It’s really quite simple.

    This pretty much sums up what I do. I really think that’s what most people actually do. I simply have higher limits for what I’m willing to spend than most people. For me, it’s about $1500 or so for graphics cards not including water blocks. It’s about $1100 for a CPU. I could have bought a Threadripper 3960X and wanted one, but $1,499.99 or whatever it is was more than I was willing to spend given that it wasn’t going to do anything for me that couldn’t be achieved by a 3950X or even a 10900K. I’m pretty good about making justifications in my mind for processors that are beyond what I actually need. After $1,100 reality sets back in and I have to give it actual thought beyond "I want that."

    it’s been the same ever since Intel increased the price of it’s HEDT chips to $1,599 for the Core i7-6950X and then all the way up to 2K for the 7980XE. Those were simply beyond what I was willing to spend. The 6950X more because it was a bad buy since I already had a Core i7-5960X@4.5GHz. On graphics cards, I’ve spent upwards of $3,000 for a pair of Titan X’s. So, $1,500 or so for an RTX 3090 doesn’t phase me in the least.

  21. Sure one thing is performance per dollar vs price point comparison. Price point comparison is very relevant of course.
  22. Sure one thing is performance per dollar vs price point comparison. Price point comparison is very relevant of course.

    I never said it wasn’t. That said, I think people ballpark what they are willing to spend and look at what aligns with that from each company. For example: If you only have $550 to spend on a GPU, then you aren’t going to really look at 3080’s. You are going to wait for the RTX 3070 and whatever AMD has to offer at the same price point.

  23. I’m just weird, I wake up and decide I feel like upgrading, then I find whatever makes me feel like I’m getting a deal and buy it :). Over the summer I asked my son if his PC was still keeping up ok… he said yup, no issues. I promptly told him we needed to upgrade it anyways, the 3300x recently launched and it was faster than his 6600k, so it made sense to pair it with a cheap board. I found a B550 board (that was a PITA at this time) that had all the features I wanted for about twice what I thought I was going to spend, and 3300x was MIA, 3600 was around MSRP, 3600x was as well… 3700x was only $260 (Amazon Prime), twice what I wanted to pay, but much better deal, then found some decent ram as well since his was 2400MHz! So, I started out wanting to upgrade for like $200-250 total, ended up spending like $500. Two weeks later I looked at my lowly R5 1600 that I basically just use to post on the forums here and do a little bit of development work and every once in a while fire up a game and decided I to needed a 3700x for no good reason :). At this time, AMD was offering AC: Valhalla for free with purchase, and that’s a game I would have bought anyways, so $250 for the CPU (thank’s Newegg), so I ended up with a second 3700x for like $190 (in my head since $250-$60 I would spend anyways)!!! So, I mostly don’t start with a budget or any real direction and buy on a whim, lol. I’m the ultimate impulse buyer, you guys and your planning and budgeting, where’s the spontaneity!

    ps. I wish I was joking, this is literally how my last 2 upgrades went. Kinda like how I ended up with a Vega 64 while looking for an rx 570… but I think my story was long enough as is!!

  24. If I needed to feel like I was getting a deal, I wouldn’t ever buy anything. Nothing is a deal in the PC world as fast as this **** depreciates.
  25. If I needed to feel like I was getting a deal, I wouldn’t ever buy anything. Nothing is a deal in the PC world as fast as this **** depreciates.

    I guess, I felt like a brand new 3700x was a pretty good deal @ what amounted to $190. If you feel that’s not a deal, then that’s fine too ;). Not everyone is patient and impulsive at the same time, lol. I’m impulsive as when I find what I feel is a good deal I’ll jump on it even if it’s not what I was thinking about getting, but I will patiently wait for said deal and not just buy something because it’s new or just came out. This is why I was on a 6600k and 1600 until this summer, I haven’t felt like anything was much of a deal when I looked, but once I decided to upgrade, it was pretty much based on what the best deal was, I’d of been perfectly fine with a 3300x but I’m perfectly fine with the 3700x too.
    I guess thinking something is a deal, is no the same as an investment. I don’t expect to make money off of it, but I do expect to spend $ that last for a certain period of time. A 3300x would have cost 1/2 the price and probably would have lasted about 1/2 as long. Nothings perfect, and like I said, I’m probably an exception to the norm, but I do feel there are deals to be had, I’ve run into them plenty of times, some better than others, but paying over MSRP is not a thing for me except one time when I needed a GPU during the mining craze, I paid $10 over MSRP for a RX 560 for my daughters birthday, which at the time was actually a good deal considering they were typically 50%-100% markups, but it took me some time to wait/find.

Leave a comment