Sponsor: Montech HyperFlow 360 Cooler on Amazon https://geni.us/dWBIbF6
AMD's Ryzen 9 9950X3D CPU launches on March 12, 2025 with an MSRP of $700, accompanied by the 9900X3D at $600 MSRP. We're reviewing and benchmarking the AMD 9950X3D CPU vs. alternatives, like the 9800X3D, Intel 285K (lol), 14900K, and base 9950X. Testing includes benchmarks in gaming and production, such as file compression, decompression, Blender CPU rendering, Adobe Premiere, and more. We also look briefly at power consumption and efficiency of the 9950X3D Zen 5 CPU for AM5 boards. Our team has prepared a number of videos to go up over the next few weeks, but we're taking a production break now that we're through all the major component reviews for the past few weeks, so we'll likely skip the 9900X3D review (but may add it to charts at some point).
Watch our 9800X3D review for more of the basics of these CPUs: https://www.youtube.com/watch?v=s-lFgbzU3LY
And our original 9800X3D coverage for more of the architecture and specs information: https://www.youtube.com/watch?v=amQOwllDXWE
Living test bench doc: https://gamersnexus.net/features/living-doc-current-test-bench-hardware-list-methodologies
The best way to support our work is through our store: https://store.gamersnexus.net/
Like our content? Please consider becoming our Patron to support us: http://www.patreon.com/gamersnexus
RELATED PRODUCTS [Affiliate Links]
AMD R9 9950X3D on Amazon: https://geni.us/3jtKVH
AMD R7 9800X3D on Amazon: https://geni.us/CJNcu4
AMD Ryzen 7 7800X3D on Amazon: https://geni.us/8YGIb
Intel 265K on Amazon: https://geni.us/sGu8f
Intel Ultra 5 245K on Amazon: https://geni.us/2dAyEN
Intel Ultra 9 285K on Amazon: https://geni.us/83NRGO
AMD Ryzen 9 9950X on Amazon: https://geni.us/GkQrF9s
Intel i7-14700K on Amazon: https://geni.us/GRUi8Ua
TIMESTAMPS
00:00 - AMD Ryzen 9 9950X3D CPU Review
03:25 - AMD R9 9950X3D CPU Specs vs 9950X
05:49 - Testing Basics and Expectations
07:57 - Frequency Validation: Blender All-Core
09:00 - Frequency Validation: Cinebench 1T
09:29 - Baldur's Gate 3 CPU Benchmarks
11:00 - Stellaris Simulation Time Benchmark
12:07 - Dragon's Dogma 2 Best CPUs
13:35 - FFXIV Dawntrail 9950X3D vs 9800X3D
15:08 - FFXIV 1440p Benchmarks
15:48 - Starfield CPU Comparison 2025
16:55 - Cyberpunk Phantom Liberty CPU Benchmarks
17:36 - F1 24 1080p
18:14 - F1 24 1440p
18:34 - Blender CPU Rendering Benchmarks
20:00 - Chromium Code Compile Test
20:47 - File Compression Benchmarks
21:41 - File Decompression 2025 CPU Benchmarks
22:13 - Adobe Premiere CPU Comparison
23:00 - Power and Efficiency: Starfield
24:01 - Power and Efficiency: 7-Zip Compression
24:52 - Power and Efficiency: 7-Zip Decompression
25:10 - Conclusion
** Please like, comment, and subscribe for more! **
Links to Amazon and Newegg are typically monetized on our channel (affiliate links) and may return a commission of sales to us from the retailer. This is unrelated to the product manufacturer. Any advertisements or sponsorshi
AMD's Ryzen 9 9950X3D CPU launches on March 12, 2025 with an MSRP of $700, accompanied by the 9900X3D at $600 MSRP. We're reviewing and benchmarking the AMD 9950X3D CPU vs. alternatives, like the 9800X3D, Intel 285K (lol), 14900K, and base 9950X. Testing includes benchmarks in gaming and production, such as file compression, decompression, Blender CPU rendering, Adobe Premiere, and more. We also look briefly at power consumption and efficiency of the 9950X3D Zen 5 CPU for AM5 boards. Our team has prepared a number of videos to go up over the next few weeks, but we're taking a production break now that we're through all the major component reviews for the past few weeks, so we'll likely skip the 9900X3D review (but may add it to charts at some point).
Watch our 9800X3D review for more of the basics of these CPUs: https://www.youtube.com/watch?v=s-lFgbzU3LY
And our original 9800X3D coverage for more of the architecture and specs information: https://www.youtube.com/watch?v=amQOwllDXWE
Living test bench doc: https://gamersnexus.net/features/living-doc-current-test-bench-hardware-list-methodologies
The best way to support our work is through our store: https://store.gamersnexus.net/
Like our content? Please consider becoming our Patron to support us: http://www.patreon.com/gamersnexus
RELATED PRODUCTS [Affiliate Links]
AMD R9 9950X3D on Amazon: https://geni.us/3jtKVH
AMD R7 9800X3D on Amazon: https://geni.us/CJNcu4
AMD Ryzen 7 7800X3D on Amazon: https://geni.us/8YGIb
Intel 265K on Amazon: https://geni.us/sGu8f
Intel Ultra 5 245K on Amazon: https://geni.us/2dAyEN
Intel Ultra 9 285K on Amazon: https://geni.us/83NRGO
AMD Ryzen 9 9950X on Amazon: https://geni.us/GkQrF9s
Intel i7-14700K on Amazon: https://geni.us/GRUi8Ua
TIMESTAMPS
00:00 - AMD Ryzen 9 9950X3D CPU Review
03:25 - AMD R9 9950X3D CPU Specs vs 9950X
05:49 - Testing Basics and Expectations
07:57 - Frequency Validation: Blender All-Core
09:00 - Frequency Validation: Cinebench 1T
09:29 - Baldur's Gate 3 CPU Benchmarks
11:00 - Stellaris Simulation Time Benchmark
12:07 - Dragon's Dogma 2 Best CPUs
13:35 - FFXIV Dawntrail 9950X3D vs 9800X3D
15:08 - FFXIV 1440p Benchmarks
15:48 - Starfield CPU Comparison 2025
16:55 - Cyberpunk Phantom Liberty CPU Benchmarks
17:36 - F1 24 1080p
18:14 - F1 24 1440p
18:34 - Blender CPU Rendering Benchmarks
20:00 - Chromium Code Compile Test
20:47 - File Compression Benchmarks
21:41 - File Decompression 2025 CPU Benchmarks
22:13 - Adobe Premiere CPU Comparison
23:00 - Power and Efficiency: Starfield
24:01 - Power and Efficiency: 7-Zip Compression
24:52 - Power and Efficiency: 7-Zip Decompression
25:10 - Conclusion
** Please like, comment, and subscribe for more! **
Links to Amazon and Newegg are typically monetized on our channel (affiliate links) and may return a commission of sales to us from the retailer. This is unrelated to the product manufacturer. Any advertisements or sponsorshi
Category
🎮️
GamingTranscript
00:00It's been a really long couple weeks with the GPU reviews and all of the chaos following
00:05it, but the quick version up front today for the 9950X3D review.
00:09The 9950X3D is comparable to the 9800X3D in gaming situations, not a big surprise there.
00:14They are basically identical in a lot of situations.
00:17In production workloads that we test, it is similar to a 9950X, it's for the most part
00:22indistinguishable for a lot of these use cases.
00:25The biggest change has been to the setup, where historically the dual CCD parts with
00:30a single CCD that's faster and a single CCD that has the extra cache, so between the two
00:36they're constructed differently.
00:38Historically it's taken some care in setup and configuration and AMD says that it is
00:43trying to solve that and it has started to do so with the chipset drivers that have come
00:48out, well actually before this, but that are really starting to roll out and see effect
00:52with the 9950X3D.
00:54That's the biggest change, it's really just to the setup.
00:56Just as an example, if you had a 7600X previously and you wanted to upgrade to a 9950X3D, our
01:02general advice would have been to just do a new Windows install, there are ways to do
01:06it without that, but that was the easiest guaranteed way of making sure you set it up
01:10right, get the drivers to hook properly, the chipset drivers to behave properly and install
01:13the things they need, and that should be fixed now in theory.
01:17Now for us, we're still isolating our environments between the core parking drives and the non-core
01:23parking drives and we've explained all that stuff in the past.
01:26So today, AMD is launching the R9 9950X3D, they say that should be available tomorrow.
01:31This is a 16-core 32-thread part, it has a listed MSRP of $700, the $600 MSRP 9950X3D
01:38will be launched alongside it, but it wasn't sampled to media.
01:41Normally that's not a good sign, we're actually probably going to skip the 9950X3D as well
01:45just because it lines up poorly with me taking a break, so we're probably skipping that one.
01:52But maybe we'll come back for it, or at least just throw it on a chart at some point, probably
01:56won't do a standalone review.
01:57Anyway, not a great sign.
01:58But as we all know, MSRP doesn't typically hold at launch.
02:01We just posted an entire video on it, it's like 42 minutes long, but it's really interesting
02:05if you want to look into that.
02:06So to keep this really simple, currently we would not pay over MSRP for a 9950X3D because
02:15it's really straightforward.
02:16The 9800X3D is cheaper, so if you're just gaming, you buy that one.
02:20If you're doing production stuff, I mean, the 9950X is $545 commonly, the 9800X3D is
02:27finally back down to $480, at least when we checked just before this, and there's some
02:31competition around too, so wouldn't pay over MSRP for it.
02:35It's that simple.
02:36If it's over MSRP, then our advice is to just ignore it and just either wait or buy something
02:41else.
02:42So today we're reviewing the 9950X3D, we're going to go through gaming tests, production
02:45benchmarks, and it's been a really long review cycle, it's been three months long now, actually
02:49four if you count Intel in December, and so we're going to keep this one really simple
02:54and just purely focused on the numbers and getting the charts out there to you.
02:58Before that, this video is brought to you by the Montech HyperFlow series of liquid
03:00coolers.
03:01The Montech HyperFlow aims to compete with other affordable liquid coolers by fighting
03:05on price to performance.
03:07In our testing, cooler performance was comparable to others in its class.
03:11The ARGB HyperFlow has an illuminated pump block with RGB LED fans that have quality
03:16rubber bumpers for vibration damping.
03:18The cooler has a six-year warranty and supports all modern Intel and AMD sockets, and you
03:23can learn more at the link in the description below.
03:25Let's get straight into it today.
03:26We'll start with the specs.
03:27The AMD 9950X3D is part of the Zen 5 architecture that launched with the 9700X and other CPUs
03:32last year.
03:33The 9800X3D swooped in a little later and cleaned up what was otherwise previously a
03:38confusing and messy launch.
03:39The 9800X3D largely made major moves for gaming CPUs and it gave us something to be actually
03:44excited about.
03:46The 9950X3D is a 16-core, 32-thread CPU with a 5.7GHz max advertised boost, 4.3GHz base
03:59clock, and L3 cache at 128MB.
04:02TDP target is 170W.
04:05For comparison, the normal 9950X also has a max advertised boost of 5.7GHz, also has
04:12a base of 4.3GHz, and it also has a TDP of 170W.
04:16These are all shared.
04:17The cache changes, of course, the 3D part, at 64MB of L3.
04:21In terms of construction, the 9800X3D is really simple.
04:24It's a single CCD plus the IO die.
04:26And so that single CCD gets the extra cache, the B-cache as they call it.
04:31They flipped how this is architected where now this was a major change that was made
04:35ground up at AMD.
04:36We have a whole separate piece on it, and I'll link that below if you want to check
04:40that out and learn about how it's all built.
04:42But now they flipped it so that the V-cache is on the bottom.
04:47It's closer to the substrate, meaning that the cores are now pushed up above it.
04:52There's a little bit less structural silicon, which helps with thermals.
04:56It gets rid of a massive just insulator, and it pushes the cores closer to the IHS, which
05:01should help them get the heat up and out without having to traverse through V-cache and additional
05:06structural silicon.
05:07So that was a major change they made with the 9800X3D.
05:10It was built ground up with the architecture itself, with Zen 5.
05:13It couldn't just be done, as AMD told us, with Zen 4 after the fact, after the architecture
05:18was done.
05:19So that was a big change that helped a lot with thermals in the 9800X3D review, and that
05:23persists with the 9950X3D and the 9900X3D.
05:27So basically everything you can learn about this architecture and X3D for this generation,
05:33you can learn it from all the many 9800X3D coverage pieces from multiple people.
05:37In the basic supply here, it's just now there's two CCDs.
05:41One of them has the extra cache, so that'd be the X3D die, and then the other one is
05:46sort of quote unquote normal in a sense, and then it's got the IO die, of course, as well.
05:49We're keeping this incredibly simple for testing this time.
05:51As always, you can find our test bench information published on the website under the LivingDock
05:55test bench page.
05:56We'll link that below.
05:57There's no third-party ads.
05:59If you want to check it out, it's pretty fast.
06:00For gaming tests, we have all new data, including the latest Windows updates and microcode updates
06:06for everything.
06:07That means we've refreshed the data set and wiped out what we had last time.
06:10So every CPU you see run has been done in the last three full days or so, so it came
06:15through really fast because of the GPUs.
06:17We got the important ones in there for production.
06:20We were able to salvage a lot of data because it hadn't changed, so we've left some of that
06:24on the charts.
06:25Otherwise, we've been completely buried by one GPU after another in an onslaught of benchmarks
06:29and follow-ups in the last week.
06:31So for this one, we're sticking to the basics.
06:33I really want to set the right expectation here.
06:35So a couple things.
06:36First of all, we don't have as many charts as normally.
06:39It's enough to make a decision.
06:40You've got all the basics for gaming and for production, and hopefully that kind of gets
06:45you the decision that you need to make.
06:47And the reason we've slimmed down on the charts is it's really simple.
06:51We are tired, and I think you'll see this repeated from a lot of reviewers right now.
06:56So I love it.
06:57It is the most fun that I can have is during these crazy review cycles, but it's hard to
07:03keep up.
07:04So anyway, a little lighter on the charts.
07:06Secondly, I am going to Taiwan in a couple of days for a mix of work and also time off
07:13for once.
07:14And so we're also going to have some rotating time off for the team, where people are going
07:18to get to go home and have a break.
07:20And so the reason I point that out is because if there's some crazy follow-up story, the
07:249950 X3D explodes or whatever, I don't know, then we're not going to be able to get to
07:30it for a while because I'm not going to move things, I'm not going to move people's time
07:34off for that.
07:35So just setting the expectation, if there's a crazy follow-up, we might not be there to
07:39cover it unless it's a while later, sorry, but I'm not going to disrupt the flow we've
07:43got for the team because everyone's worked really hard to get to this stage.
07:46And this is the last review we're getting live before we get to do some other stuff
07:50for a little while before the next round of chaos in April with probably the 50s, 60s,
07:54and the 90s, 60s.
07:55Anyway, enough about us.
07:56Let's get into the data.
07:57Frequency analysis up first.
07:59We do this testing to ensure the CPUs are functioning as expected and to help explain
08:03the performance later.
08:04First up is the Blender all-core workload.
08:06In this test, the 9950 X3D had a frequency plot that started at about 50-50MHz but settled
08:12closer to 50-20MHz to 50-80MHz during testing.
08:16This chart is intentionally zoomed in to make it easier to see, so the scale purposefully
08:20does not start at zero.
08:22For comparison, the 9950 X non-3D had higher peaks but similar valleys.
08:26It ranged from 50-10 to 50-80MHz.
08:29In terms of the average frequency over the course of the test, the 9950 X3D averaged
08:3450-38MHz all-core to the 9950 X's 50-36MHz, about the same there, but the X3D CPU did
08:42so with fewer peaks and more level frequencies in the middle of its range.
08:46We think this will be beneficial to it in gaming.
08:48The 9800 X3D's 50-20 to 50-25MHz all-core put it well above both.
08:54This will help it in some specific workloads, but obviously the lower core count will set
08:58it back elsewhere, just depends.
09:00The next chart is for frequency in a Cinebench single-threaded workload.
09:03This has the 9950 X3D up in the range of 5650 to 5725MHz, which hits AMD's advertised
09:09frequency of 5.7GHz.
09:12The 9950 X holds its frequency steadier and with fewer dips between tile cycles, but is
09:17overall comparable.
09:18The 9800 X3D holds 50-25MHz throughout the test, so it's lower than both when in a
09:24single-threaded workload in this situation.
09:27Let's get into some of the games.
09:28Ballistic 8.3 is up now, and this one had the AMD R9 9950 X3D at 155FPS average, technically
09:34becoming a new chart topper.
09:35The 9800 X3D was our chart topper last time, and is now functionally tied with the 9950
09:39X3D as the best CPU on the chart.
09:41The good news is that the 9950 X3D doesn't appear to be suffering from its dual-CCD approach,
09:46so parking is functioning properly and the CPU is not hamstrung by its extra threads
09:49this time.
09:50The 9950 X3D outranks the 7950 X3D by similar margins as the 9800 X3D did, it's 23% higher
09:56average framerate, with lows comparable for the average.
09:59The 7950 X3D outdid the 7950 X, which we're using old data for for that one, but should
10:04be no greater than 2-3% different based on our study of this test, and the amount that
10:08the X3D 7950 is outperforming it by is 29%.
10:11That's with a proper setup of the 7950 X3D this time.
10:14As compared to the 9950 X at 101FPS average, the 9950 X3D outdid it by 54%.
10:20The 9950 X is closer to the 7950 X for performance, which makes sense.
10:24This game really benefits from the extra cache, it looks like.
10:27In fact, an easy example of this is the 5700 X3D versus the 5600 X3D.
10:32In some games, the 5600 X3D outperforms the 5700 X3D because of its higher clock rate.
10:38In this instance, though, the cache on the core counts more beneficial than the frequency.
10:42The 5800 X3D remains an excellent CPU, up at 120FPS average or so.
10:46The 9950 X3D and 9800 X3D outrank it by about 30%.
10:50As for Intel, it remains uncompetitive here.
10:53The 285K is getting crushed by two prior Intel generations for reasons discussed in
10:57that review, and that's with new Windows updates.
11:00Stellaris is one of our favorite CPU benchmarks because it looks at time rather than framerate,
11:04which is the most tangible to a user and most directly influenced by the CPU.
11:08Players of 4X or other grand strategy games like Total War with a campaign map, Galactic
11:12Civilizations 4 with turn pacing, that's a great game by the way if you haven't played
11:16it you should, and Civilization would all see value here.
11:19For Stellaris, the 9800 X3D and 9950 X3D both perform at the top of the chart.
11:24The 9800 X3D outperformed the 9950 X3D with a reduction in simulation time of 5%, that's
11:30near error, but not quite.
11:32This seems to be a combination of a higher base clock and utilization.
11:35The 9950 X3D is definitely working as expected, though, because it's outperforming the 9950X
11:40significantly.
11:41The simulation time requirement drops by almost 15%, from 32.3 seconds to 27.6 seconds.
11:47This is the one game where Zen 5 in particular had stronger gains over Zen 4, with the 9700X
11:53doing well here and proving that.
11:55That's from IPC Uplift overall, where Zen 5 has benefited.
11:58Intel's 285K is competitive with the 7800 X3D and 9700X at least, the 14900K and 14700K
12:04are within error of each other.
12:05In Dragon's Dogma 2, the 9950 X3D leads the chart.
12:08It landed at 132 FPS average here, passing the 9800 X3D by a measurable but irrelevant
12:14and unnoticeable to an end user, 3.2%.
12:17Both CPUs lead all of Intel's, although Intel at least lands its prior two generations ahead
12:22of the 7950 X3D and 7800 X3D with the game's updates.
12:26This game really seems to benefit from extra cash, with the 9950 X3D leading the 9950X
12:30by 46%, and the 9800 X3D leading the 9700X, although they have other differences, by 41%.
12:37Dragon's Dogma 2 remains heavy on CPUs and NPC intensive areas.
12:42The 285K continues to impress with how much of a downgrade it is from not only AMD's current
12:48generation, but Intel's past generations, and so it looks like even with Windows updates
12:53and game updates and everything else, our conclusions are more or less the same on the
12:57Ultra 200 series.
12:59We added the older results for the 3700X and 2600 to this chart, now we noticed that performance
13:03on older generations doesn't appear to have changed much, at most there might be about
13:08a 5% change here, but we don't think so.
13:10Even with that though, anything's an upgrade, we just wanted to put them on here in case
13:13it's a helpful reference.
13:15Intel has seen the most upgrades since our last round of tests in this game.
13:20This game has gotten updates, so it's possible that those were targeted at Intel, it got
13:24uplift here as a result, Windows updates could also affect it.
13:27We consistently saw uplift across Intel CPUs in this specific game, that's shifted the
13:31relative ranking of the 14th and the 13th gen against the 7800X3D.
13:35Final Fantasy XIV Dawn Trail is up now, in this one the 9800X3D ran at 380FPS average,
13:40with the 9950X3D at 373FPS average.
13:44We observed relatively wide run-to-run variance in some of these results, so the error bars
13:48are wider than typical.
13:49The 9800X3D leads the 9950X3D by just 2%, so they're functionally equal here.
13:55The 9950X3D also bests its 9950X non-3D variance by 50FPS or so here, so that's about 16% in
14:02average framerate.
14:03The 1% lows are also significantly improved, indicating that frametime pacing keeping up
14:07with improvements in the average framerate is what we're seeing here with the 9950X3D.
14:12The improvement over the 7950X3D is 5.8%, not a lot there.
14:16Intel's closest CPUs don't appear until the 14900K at 310FPS average.
14:21This is mostly interesting because there was a time when Final Fantasy's prior benchmark
14:24versions were entirely dominated by Intel, with a clean division halfway down the chart,
14:30which is actually what we're seeing now favoring AMD, relegating Intel to the bottom, so it's
14:34flipped in recent years and generations.
14:36Intel's 285K underperforms against its prior two generations.
14:39There was no change in performance against last time for the 285K.
14:43Intel's one advantage in this test is frametime pacing, where the 0.1% lows indicate that
14:48Intel's CPUs generally have more consistent frame-to-frame intervals than AMD's in this
14:52game, although not by an amount that'd change your experience in a notable way.
14:56The 5600X3D outperforms the 5700X3D in this game.
15:00This has been known and is because of higher clock speeds on the 5600X3D, which proves
15:05valuable in this game versus the extra cores.
15:08At 1440p, the top of the chart truncates as a result of GPU limitations on the 4090.
15:14We'll move to the 5090 for a full revamp of our CPU testing for the next major architecture
15:18release, but for now, this is where we cap out.
15:20We're sure this is deeply disappointing to all 12 of you who have an RTX 5090.
15:25Sorry, maybe next time.
15:27The 9800X3D and the 9950X3D are about the same, once again.
15:31The 7950X3D is also now about the same, as is the 9700X, thanks to external limitations.
15:36This is a good reminder that the gains once scaling graphics are most seen in time-based
15:41situations or in seriously heavy CPU games like Dragon's Dogma 2, but otherwise a lot
15:45of the times you'll get the most uplift from a GPU.
15:48Starfield is up next.
15:49In this one, we have the 9950X3D at a 171 FPS average, leading the 9800X3D's 165 FPS
15:56by 3%.
15:57The 9800X3D was notably ahead of the 7800X3D, and the 9950X3D continued that, though neither
16:03had as revolutionary of a gain as we've seen in some of the other benchmarks.
16:07The 14900K trails the 7800X3D, improving upon its prior round result in a meaningful way.
16:13However, because of the improvements we've seen in prior generations, the 285K now falls
16:17back behind Intel's 14900K in this test.
16:20The 285K still regresses and generally embarrasses Intel, trailing even the 13700K at times.
16:26Intel has continually tweaked its microcode on these prior generations, so it's possible
16:30that they rolled out a microcode that had lost some performance at some point and they've
16:34regained some now in the 1413 series.
16:36We've updated BIOS to the newest version for each round, so that is a change that we make
16:41each time we run these benchmarks.
16:43With the 9950X at 124FPS average, the 9950X3D improves by 37%.
16:49That's one of the larger gains.
16:50Of course, if you're not going to use the extra cores, the 9800X3D makes more sense
16:54for value.
16:55Cyberpunk 2077 is back in our CPU test suite again with the Phantom Liberty expansion.
16:59Tested at 1080p medium here, the 9800X3D and the 9950X3D both ran at about 219FPS average
17:06and were well within run-to-run variance at only fractions of a frame per second apart.
17:10If the 7800X3D trails, but not by much, it'd be roughly the same experience as these two.
17:15The lead of the 9950X3D over the 9950X is 37% again, matching some of the other games
17:21we've seen.
17:22The Intel 200 series outdoes the prior generations here sometimes, finally, with the 285K at
17:26170FPS average.
17:27Unfortunately, that's still below the AM4 5700X3D and 5600X3D, but at least it's a little
17:33better than its last-gen stuff.
17:36F124 at 1080p is up now as the last game for today.
17:39This one has the 9950X3D and 9800X3D again roughly tied, with the 7800X3D not far behind.
17:45The advantage is only 7%.
17:47The 9950X3D leads the 9950X non-3D variance by 29%, slightly reduced from the advantage
17:54seen in other games.
17:55We might be hitting a GPU limit here.
17:58Intel's 14900K is the closest competition it has, released in 2023, with the 285K from
18:04last year down at 9950X levels.
18:07The 5600X3D and 5700X3D results again show that this game likes frequency and IPC to
18:12some extent.
18:131440p is almost exactly the same at the bottom half, with the top switching around due to
18:17GPU overhead and limitations on GPU scaling.
18:20The 5800X3D falls down the ranks as the 14th and 13th gen handle the overhead a little
18:25better with more stable frametime pacing, which helps with the average.
18:28Otherwise, things are about the same sans limitations of scaling for the 9950X3D.
18:33We're moving on to production tests now.
18:35This is where the 16-core CPUs do well.
18:37Historically, we've seen the X3D variants of these CPUs underperform against the non-X3D
18:41parts due to power allocation to allow higher boosting.
18:45Extra cache doesn't help in our testing here normally, but we'll see if the 9950X3D breaks
18:51the general trend.
18:52Blender testing hasn't changed since our October round.
18:55We ran validation on several CPUs and the results came out basically identically, so
18:59we can keep a lot of data for more comparisons.
19:02This should help those of you on older hardware.
19:05The 9950X3D required 6.6 minutes to complete the render, which is about tied with the 9950X.
19:11It was technically faster, but in reality, these are tied.
19:15That's good news for the X3D part, though, because past X3D CPUs like the 7950X3D have
19:20been technically slightly slower than their non-3D variants.
19:23That's not because of scheduling or parking, but because the frequency was just slower
19:27in an all-core workload.
19:28Another good example is the 7800X3D, which was slower than the 7700X by time required.
19:34Or the 5800X3D, which also required more time than the 5800X.
19:38The 9950X3D was the first to break this trend in a big way.
19:43Technically, the 9800X3D looked like it was doing that against the 9700X, but the power
19:48target was what brought most of the change.
19:50The 9950X3D outperforms the 285K here with about a 7% reduction in total time required
19:57to complete the render.
19:58Chromium, Code Compile, and Windows is another where the dataset hasn't changed, so we
20:03were able to salvage it after validation.
20:05The 9950X3D required 81 minutes to completely compile, which is comparable to the time required
20:11for the 9950X, but technically improved.
20:14This puts it marginally ahead of the freshly retested 285K, with a reduction in time required
20:20to compile from 285K to 9950X3D of 4.7% less time required.
20:26The 4900K required 88 minutes here, with the 265K at 98 minutes.
20:32The 9950X3D is the new leader in our compile test.
20:35This is not going to be representative of every type of code compile, just like none
20:39of these tests is representative of every angle of the use case.
20:42However, the way we test it, the 9950X3D is the new leader short of going to Threadripper.
20:47Data for compression and decompression can't be salvaged, so it's all new.
20:51In the 7-zip file compression testing, the 9950X3D led the chart at 206,643 MIPS, or
20:57millions of instructions per second.
20:59That has it just ahead of the 9950X by 3.3%.
21:02This is one of the tests where cache can help, depending on its implementation.
21:07The 5600X3D and 5600X are good examples of this.
21:10The X3D part has a lower advertised frequency, but manages to still roughly tie the 5600X.
21:16The 9950X3D outperforms the 7950X3D by 8.5% here, which completed at 191,000 MIPS.
21:24The 14900K is next, at roughly 189,000 MIPS, and then the 13900K.
21:29The 285K follows all of these, down at 179,000 MIPS.
21:33Core count clearly matters in this test, the 3950X 16-core CPU is outperforming the 5900X
21:3812-core CPU and 9700X 8-core CPU.
21:42In 7-zip decompression, we measured the 9950X3D at 277,000 MIPS with the 9950X non-3D at 272,000.
21:50You wouldn't really benefit from the 9950X3D in a meaningful way in either compression
21:55or decompression on this workload, the 9950X achieves pretty much all the performance already,
21:59so you'd need use cases that more directly leverage the cache to get value out of the
22:049950X3D.
22:05Intel's 14900K is its closest competitor, followed by the 14700K, and then the 285K.
22:10We saved some of the data for Adobe Premiere as well.
22:13The biggest swing was to Intel's 12th and 14th gen CPUs here in 13th, where we saw improvement
22:18from Windows updates recently.
22:21Most other parts stayed relatively stationary, so any 12th to 14th gen CPUs that are date
22:26stamped as data prior to this round, those would move around a little bit, so be aware
22:32of that.
22:33We decided to leave those on the charts anyway because they didn't move by a ton and it might
22:38still help some of you who are on those parts, but just be aware that it's not 100% life
22:42for like, it is pretty close though.
22:43The 9950X3D scored 11,600 points in the Puget Suite Aggregate Extended Scoring for Premiere,
22:49which puts it at the top of the chart.
22:51It bests the 9950X by 5.8%, with the 14900K closest to it, and then the 285K.
22:58The improvement over the 7950X non-3D is 7%.
23:01We'll keep power and efficiency testing short this time to just show a couple situations.
23:05In Starfield, the 9950X3D ended up at 1.7 FPS per watt, putting it behind the 7950X3D
23:11and the 7800X3D, but tied with the 5700X3D and 9800X3D.
23:16The 9950X3D pulled 98.8 watts when playing this game, and Starfield is one of our games
23:21that most heavily loads the CPU, but still nothing like an all-core blender workload,
23:25and especially not when you're parking basically half of the CPU to do more or less nothing.
23:31The 9950X non-3D part pulled 168 watts in this same test, still doing core parking,
23:37but ended up pulling lower wattage, part of this is due to voltage pinning.
23:41That puts it down at 0.7 FPS per watt, and that means the 9950X pulled nearly 70 watts
23:46more than the 9950X3D, or about a 70% increase in power consumption despite running at a
23:52lower frame rate.
23:53In terms of FPS per watt, the 9950X3D is both higher frame rate and lower power, and so
23:58it's far more efficient.
24:00It's still not as efficient as the low power 7800X3D though.
24:037-Zip compression shows that the 9950X3D can still be power hungry, and in our compression
24:08efficiency testing, the 9950X3D pulled 203.8 watts.
24:12That put it at 1014 mips per watt, or millions of instructions per second per watt, which
24:17makes it less efficient than about half of our chart.
24:20The CPU is the best performer, but not for efficiency, and that's because it's pulling
24:25204 watts, so it's efficiency has decreased compared to some others.
24:29The 9950X scored 979 mips per watt, and pulled the same power at 204 watts, making it less
24:35efficient than the 9950X3D.
24:38The 7800X3D is a lower performer overall, and in big ways, but has such impressively
24:42low power consumption that it ends up being the most efficient.
24:45Of course, if you were serious about running this kind of workload all the time, you'd
24:49still want something more powerful than the 7800X3D.
24:52Decompression testing looks better for the 16-core parts, with the 7950X3D proving incredibly
24:56efficient here, followed by an impressive result from the 7950X non-3D with eco mode
25:01enabled.
25:02The 9950X3D ran at 1358 mips per watt, putting it slightly ahead of the 9950X.
25:08They're still in the middle of this chart, though.
25:10That's it, you've got the numbers.
25:11This is going to be one of those where we let you make the decision based on the numbers
25:14we've just presented.
25:15The quick version, again, is that for gaming, basically think of it like a 9800X3D.
25:19We didn't run into any major issues with the 9950X3D.
25:22That in and of itself is kind of an accomplishment for AMD.
25:24They've really struggled over the years with the dual CCDs where one has X3D with extra
25:29cache on it.
25:30Over the years, it's taken them some time to get to a place where it is not regressive
25:35and where it's a little easier to set up.
25:37And so the 9950X3D does appear to do that in our experience so far, and that is a major
25:43improvement for AMD.
25:44It's taken them some generations to get there, so good on them for that.
25:48If you are building only a gaming computer, it's really simple to us.
25:52If you've got the budget for looking at stuff this expensive, we think you should actually
25:55scale it down and go for a 9800X3D.
25:58The reason is it's just not that big of a difference.
26:00A lot of times it's kind of within error or they trade places with each other.
26:04So if it's purely gaming and you really don't explicitly know that you need those extra
26:09cores, we think you go with the 9800X3D.
26:12You save some money.
26:13Intel is out of this conversation right now.
26:14They are not a part of the high-end, expensive CPU for gaming build scenario at the moment,
26:21so they're just absent.
26:23The 9950X non-3D makes sense still if you're doing a production-heavy build where you don't
26:30have a use case that explicitly uses that extra cache.
26:33There's definitely use cases like that out there.
26:36We see a little bit of that in some of our 7-zip compression or decompression testing,
26:40but for the most part, the things we test, it doesn't tend to benefit from the extra
26:46cache in non-gaming scenarios.
26:48So really where this, like the other X3D 16-core parts, shines is a more limited use
26:55case, and there's still a lot of you out there who do this, where you have some mix of really
27:00heavy production and of really heavy gaming.
27:03So if you do stuff like a lot of compression, decompression, maybe you render things on
27:08the CPU for one reason or another.
27:10Maybe you are heavy into Premiere, as an example, or you do a lot of code compiles.
27:15We've got charts for all those.
27:17And if you then, when you're done with all that work, play games and don't want a super
27:21high frame rate, that's kind of the use case for this.
27:24If you're one camp or the other exclusively, then you can save some money by doing a 98X3D
27:29or a 9950X, and then of course Threadripper is on the other side, which we have separate
27:34reviews for that.
27:35It's not on these charts.
27:36So that's pretty much it for the fast recap.
27:38We wouldn't pay more than MSRP for this.
27:40CPUs tend to stick closer to that.
27:42They are still going to have stupid prices from either some retailers or from third party
27:47sellers on those retailers and their marketplaces.
27:50You've got all the numbers.
27:51You can make a decision and look at the value and think about if you like it or not.
27:55But for us, that is going to be it for this one.
27:58We've got a lot of content lined up, even though there's going to be a bit of a break
28:02on production for us.
28:03So a bit of a production pause where part of what we're doing is for some of the time,
28:07we're going to focus on some new testing methods, some new chart styles that I'm working on,
28:12it's going to be really fun.
28:13So we do this about four times a year.
28:15We halt production for a little bit, overhaul things and kind of get it prepped and do some
28:19fun stuff to keep it fresh in the testing and do new things.
28:24And then for the other part of that period where we're halting production, it's just
28:27going to be time off.
28:29So everyone's got a chance to recover from the big review cycle.
28:33And otherwise, we've got a ton of content lined up for it because we've been preparing
28:37for that.
28:38So it won't look too different for you publicly.
28:39It's going to be about one week where we're light on videos, and that's why.
28:42So anyway, thanks for watching.
28:43We'll have an article version of this review up pretty soon if you want to check that out
28:47on GamersAccess.net.
28:48Jimmy on the team has been converting all of these reviews very quickly over on the
28:52site if you want the written reference to have something quick to go through and all
28:55the charts.
28:56And let us know what you think of the 9950X3D.
28:59But that is, that's going to be all.
29:02Thanks for watching.
29:03Subscribe for more.
29:04Go to store.gamersaccess.net to help us out directly.
29:05We'll see you all next time.