• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

AMD Ryzen 9 5900X

Joined
Jan 31, 2011
Messages
2,016 (0.51/day)
System Name Ultima
Processor AMD Ryzen 7 5800X
Motherboard MSI Mag B550M Mortar
Cooling Arctic Liquid Freezer II 240 rev4 w/ Ryzen offset mount
Memory G.SKill Ripjaws V 2x16GB DDR4 3600
Video Card(s) Palit GeForce GTX 1070 Super JetStream 2Ghz
Storage Samsung 970 Evo Plus 500GB nvme, 1TB Crucial MX500 SSD sata, 1TB Seagate Barracuda 7200rpm HDD sata
Display(s) LG 23MP68VQ-P IPS 75HZ 23" 1080p MONITOR
Case DarkFlash DLM21 Mesh
Audio Device(s) Onboard Realtek ALC1200 Audio/Nvidia HD Audio
Power Supply Corsair RM650
Mouse Logitech G304 | Wacom Intuos CTH-480
Keyboard A4Tech B314 Keyboard
Software Windows 10 Pro
i had a good deal with a 3600Mhz CL18, i wonder if it's still fine, i can put it on CL16 still i guess
 
Joined
Jan 27, 2019
Messages
85 (0.08/day)
Theree's Many things wrong with this review, really disappointing to see this from TPU.. TPU disappointment strikes yet again. I dont believe your 65c load temps on 10900k Its fake for sure.
 
Joined
Feb 13, 2017
Messages
58 (0.03/day)
I love TPU for the design / graphics presentation, but this review really dropped the ball. Sorry, 3600C16 memory is dirty cheap now (we're at DDR4 peak, DDR5 next year), and AMD said that this platform could go to 4000 Mhz 1:1 mem/flck, so this piece really feels biased / tuned to protect Intel, even if undesired. Buying a ~550 CPU and saving $10 bucks on memory - for a staggering performance loss - makes no sense.
 
Joined
Jan 27, 2019
Messages
85 (0.08/day)
I love TPU for the design / graphics presentation, but this review really dropped the ball. Sorry, 3600C16 memory is dirty cheap now (we're at DDR4 peak, DDR5 next year), and AMD said that this platform could go to 4000 Mhz 1:1 mem/flck, so this piece really feels biased / tuned to protect Intel, even if undesired. Buying a ~550 CPU and saving $10 bucks on memory - for a staggering performance loss - makes no sense.

They also used a 2080Ti and their CPU load temps just seem so wrong..
 
Joined
Oct 10, 2018
Messages
865 (0.75/day)
I'm very disappointed with some people coming on here, trying to accuse, mislead, and throw a mud at integrity of these tests and testers themselves. If you dont like it, bugget off! Now what I noticed is this time around AMD played safe with boost clocks, in most cases boost actually exceeded their advertised numbers. Remember last year? They could not even get close to those numbers ands on top of that no issues on bios/software. This was the biggest question mark for AMD, and I think today they passed it. Well done AMD.
 

Yosar

New Member
Joined
Apr 16, 2019
Messages
4 (0.00/day)
I call bullshit on these tests. Witcher 3 is especially bullshit.


In every test 5900X beats 10900K easily in this game. Not even close to be considered equal.
All those sites are wrong and only TechpowerUp is right? Sorry, but this revierw is so much dishonest that I stoped belive in any review from this site.
Reviewer fake those tests or doesn't know what he does. And I don't know what is worse, dishonesty or incompetence.
Oh, and any multiplayer or strategic game totally ignored. Worthless review by all means.
 
Joined
Jan 27, 2015
Messages
1,139 (0.45/day)
System Name Legion
Processor i9-10850K
Motherboard Asus Prime Z490M Plus
Cooling Arctic Liquid Freezer 2 240mm AIO
Memory G.Skill Ripjaws V 32GB (2 x 16GB) DDR4-3200 F4-3200C16D-32GVK
Video Card(s) EVGA GeForce RTX 2060 KO Ultra
Storage Team Group MP33 Pro 2TB M.2 2280 PCIe NVMe 3.0 x4 + WD Blue 1TB SATA SSD
Display(s) Acer K272HUL 1440p / 34" MSI MAG341CQ 3440x1440
Case Lian Li 205M
Power Supply PowerSpec 650W 80+ Bronze Semi-Modular PS 650BSM
Mouse Logitech MX Anywhere 25
Keyboard Logitech MX Keys
Software Lots
Anands used 2933 on their 10900K, and 3200 on their Zen 3 CPUs. No sane person with a 10700K or above is going to run DDR4-2933.

So they basically crippled the Intel systems.

AnandTestSetups.JPG



This is the effect of using higher speed RAM with Intel chips. It's like 5% going to DDR4-3200, and 8% using DDR4-3600 :

RAMOC.JPG
 
Joined
Dec 31, 2009
Messages
19,324 (4.43/day)
Benchmark Scores Faster than yours... I'd bet on it. :)
Anands used 2933 on their 10900K, and 3200 on their Zen 3 CPUs. No sane person with a 10700K or above is going to run DDR4-2933.

So they basically crippled the Intel systems.
This is a like vs like situation. Both values are the maximum rated for the cpu. Anything above that is considered overclocking (the imc). To be fair (imo) they should be tested at the same speeds. 3600 for example, is overclocking both imcs and lets each cpu stretch its legs.

You also need to understand the majority don't overclock in the first place.
 
Joined
Jan 27, 2015
Messages
1,139 (0.45/day)
System Name Legion
Processor i9-10850K
Motherboard Asus Prime Z490M Plus
Cooling Arctic Liquid Freezer 2 240mm AIO
Memory G.Skill Ripjaws V 32GB (2 x 16GB) DDR4-3200 F4-3200C16D-32GVK
Video Card(s) EVGA GeForce RTX 2060 KO Ultra
Storage Team Group MP33 Pro 2TB M.2 2280 PCIe NVMe 3.0 x4 + WD Blue 1TB SATA SSD
Display(s) Acer K272HUL 1440p / 34" MSI MAG341CQ 3440x1440
Case Lian Li 205M
Power Supply PowerSpec 650W 80+ Bronze Semi-Modular PS 650BSM
Mouse Logitech MX Anywhere 25
Keyboard Logitech MX Keys
Software Lots
I call bullshit on these tests. Witcher 3 is especially bullshit.


In every test 5900X beats 10900K easily in this game. Not even close to be considered equal.
All those sites are wrong and only TechpowerUp is right? Sorry, but this revierw is so much dishonest that I stoped belive in any review from this site.
Reviewer fake those tests or doesn't know what he does. And I don't know what is worse, dishonesty or incompetence.
Oh, and any multiplayer or strategic game totally ignored. Worthless review by all means.


Another one who can't read a graph.

Tell us, which chips from that eurogamer link won this. I'll give you a hint, it starts with a 10 :

eurogamer1.jpg


Or this :

eurogamer2.jpg


Or this :

Eurogamer3.JPG
 
Joined
Mar 31, 2012
Messages
822 (0.23/day)
Location
NL
System Name SIGSEGV
Processor INTEL i7-7700K | AMD Ryzen 2700X
Motherboard QUANTA | ASUS Crosshair VII Hero
Cooling Air cooling 4 heatpipes | Corsair H115i | Noctua NF-A14 IndustrialPPC Fan 3000RPM
Memory Micron 16 Gb DDR4 2400 | GSkill Ripjaws 32Gb DDR4 3200 3400(OC) 14-14-14-34 @1.38v
Video Card(s) Nvidia 1060 6GB | Gigabyte 1080Ti Aorus
Storage 1TB 7200/256 SSD PCIE | ~ TB | 970 Evo
Display(s) 15,5" / 27"
Case Black & Grey | Phanteks P400S
Audio Device(s) Realtek
Power Supply Li Battery | Seasonic Focus Gold 750W
Mouse g402
Keyboard Leopold|Ducky
Software LinuxMint KDE |UBUNTU | Windows 10 PRO
Benchmark Scores i dont care about scores
Joined
Jan 27, 2015
Messages
1,139 (0.45/day)
System Name Legion
Processor i9-10850K
Motherboard Asus Prime Z490M Plus
Cooling Arctic Liquid Freezer 2 240mm AIO
Memory G.Skill Ripjaws V 32GB (2 x 16GB) DDR4-3200 F4-3200C16D-32GVK
Video Card(s) EVGA GeForce RTX 2060 KO Ultra
Storage Team Group MP33 Pro 2TB M.2 2280 PCIe NVMe 3.0 x4 + WD Blue 1TB SATA SSD
Display(s) Acer K272HUL 1440p / 34" MSI MAG341CQ 3440x1440
Case Lian Li 205M
Power Supply PowerSpec 650W 80+ Bronze Semi-Modular PS 650BSM
Mouse Logitech MX Anywhere 25
Keyboard Logitech MX Keys
Software Lots
This is a like vs like situation. Both values are the maximum rated for the cpu. Anything above that is considered overclocking (the imc). To be fair (imo) they should be tested at the same speeds. 3600 for example, is overclocking both imcs and lets each cpu stretch its legs.

You also need to understand the majority don't overclock in the first place.

I think it's valid either way, as long as the viewer knows it . Out here in enthusiast space though, almost no-one buys anything slower than DDR4-3200, in fact almost everyone has DDR4-3200 if you peruse system specs. If you are talking about OEM systems, then they should probably test with DDR4-2666 on both platforms, because that is by a wide margin the predominant speed of RAM that will be put on OEM boxes.
 
Joined
Dec 31, 2009
Messages
19,324 (4.43/day)
Benchmark Scores Faster than yours... I'd bet on it. :)
I think it's valid either way, as long as the viewer knows it . Out here in enthusiast space though, almost no-one buys anything slower than DDR4-3200, in fact almost everyone has DDR4-3200 if you peruse system specs. If you are talking about OEM systems, then they should probably test with DDR4-2666 on both platforms, because that is by a wide margin the predominant speed of RAM that will be put on OEM boxes.
Enthusiast land is but a drop in the bucket my friend. Not much would have changed if they ran intel at 2933 compared to 3200 (their lax rated). Again, most aren't overclocking. Overclocking is a variable metric depending on how far users want to push things.

In short, let it go... that isn't a great argument you brought up with that anand review. ;)
 
Joined
May 3, 2018
Messages
545 (0.41/day)
Wow how bad does the 1800X look nowadays. I only have a lowly 1700X and it's amazing to see how even the 5600X beats out the 8 cores designs in a lot of tests. I have two PCs and one is a dinosaur running Intel 3570K and 1070 and other is 1700X + 1080 Ti. I was just going to update the old one, make that my powerhouse and the 1700X machine becomes the secondary PC, but now I will even update that one. One build will be full on, 5900X + 6800XT, the other I'll just do cpu, MB, memory and GPU, probably 5600X + either keep the 1080Ti or get 6800.
 
Joined
Jan 27, 2015
Messages
1,139 (0.45/day)
System Name Legion
Processor i9-10850K
Motherboard Asus Prime Z490M Plus
Cooling Arctic Liquid Freezer 2 240mm AIO
Memory G.Skill Ripjaws V 32GB (2 x 16GB) DDR4-3200 F4-3200C16D-32GVK
Video Card(s) EVGA GeForce RTX 2060 KO Ultra
Storage Team Group MP33 Pro 2TB M.2 2280 PCIe NVMe 3.0 x4 + WD Blue 1TB SATA SSD
Display(s) Acer K272HUL 1440p / 34" MSI MAG341CQ 3440x1440
Case Lian Li 205M
Power Supply PowerSpec 650W 80+ Bronze Semi-Modular PS 650BSM
Mouse Logitech MX Anywhere 25
Keyboard Logitech MX Keys
Software Lots
Enthusiast land is but a drop in the bucket my friend. Not much would have changed if they ran intel at 2933 compared to 3200 (their lax rated). Again, most aren't overclocking. Overclocking is a variable metric depending on how far users want to push things.

In short, let it go... that isn't a great argument you brought up with that anand review. ;)


But it is where the difference came from. People are talking about 0-5% difference in the reviews. Look at the 9900K at 2666 vs 3200, that's 14.4% to be precise. 3200->3600 is 1%. Those margins are more than enough for memory to explain this if AT ran their Intel chips at 2933 and AMD chips at 3200.

As far as OEM arguments, if it's an OEM type setup config they should be running at typical OEM speeds which would be 2133 - 2933 at best. For enthusiast / custom builds, it's almost all 3200. If they want to max out the performance, use the fastest you can get stable.

Honestly for anyone thinking about an OEM system (Dell, HP, etc), *all* of these enthusiast review sites are garbage information.

Edit: added the chart showing effects of RAM on 9900K vs 3900X I was referring to.

RAMOC.JPG
 
Last edited:
Joined
Dec 31, 2009
Messages
19,324 (4.43/day)
Benchmark Scores Faster than yours... I'd bet on it. :)
But it is where the difference came from. People are talking about 0-5% difference in the reviews. Look at the 9900K at 2666 vs 3200, that's 14.4% to be precise. 3200->3600 is 1%. Those margins are more than enough for memory to explain this if AT ran their Intel chips at 2933 and AMD chips at 3200.

As far as OEM arguments, if it's an OEM type setup config they should be running at typical OEM speeds which would be 2133 - 2933 at best. For enthusiast / custom builds, it's almost all 3200. If they want to max out the performance, use the fastest you can get stable.

Honestly for anyone thinking about an OEM system (Dell, HP, etc), *all* of these enthusiast review sites are garbage information.

Edit: added the chart showing effects of RAM on 9900K vs 3900X I was referring to.

Brotha... we can't help how they spec their systems out. 2933 is the max for intel, 3200 for AMD. I have no idea what foreign language site you're quoting, but you should provide a link so we can see exactly what game that is and have some context for this image you keep posting. Also note that 2666 JEDEC is below the max... and JEDEC timings are extremely high... nothing of which Anandtech did, note. Their 2933 was with more appropriate than JEDEC timings. Same with the AMD.

Here are some links for you to chew on... notice here there isn't much of a difference in games.
...and intel

I'm sure there are some results like this that show huge gains, but those are anomolous for the most part. Memory speeds, generally yield a couple % gains. AMD tends to benefit more, especially with lower latency. Again, give it up... whatever you're arguing here. :)
 
Joined
Jan 27, 2015
Messages
1,139 (0.45/day)
System Name Legion
Processor i9-10850K
Motherboard Asus Prime Z490M Plus
Cooling Arctic Liquid Freezer 2 240mm AIO
Memory G.Skill Ripjaws V 32GB (2 x 16GB) DDR4-3200 F4-3200C16D-32GVK
Video Card(s) EVGA GeForce RTX 2060 KO Ultra
Storage Team Group MP33 Pro 2TB M.2 2280 PCIe NVMe 3.0 x4 + WD Blue 1TB SATA SSD
Display(s) Acer K272HUL 1440p / 34" MSI MAG341CQ 3440x1440
Case Lian Li 205M
Power Supply PowerSpec 650W 80+ Bronze Semi-Modular PS 650BSM
Mouse Logitech MX Anywhere 25
Keyboard Logitech MX Keys
Software Lots
It's computerbase.de - a very active site. I posted the link the first time I put up that image. Most of the recent 'revelation' that in fact Intel does scale with RAM speed - better than Ryzen in fact - started from this review.


Brotha... we can't help how they spec their systems out. 2933 is the max for intel, 3200 for AMD. I have no idea what foreign language site you're quoting, but you should provide a link so we can see exactly what game that is and have some context for this image you keep posting. Also note that 2666 JEDEC is below the max... and JEDEC timings are extremely high... nothing of which Anandtech did, note. Their 2933 was with more appropriate than JEDEC timings. Same with the AMD.

Here are some links for you to chew on... notice here there isn't much of a difference in games.
...and intel

I'm sure there are some results like this that show huge gains, but those are anomolous for the most part. Memory speeds, generally yield a couple % gains. AMD tends to benefit more, especially with lower latency. Again, give it up... whatever you're arguing here. :)

#1 Their test is with a 9900K, which was Intel Spec'd for 2666.

If you truly mean what you say about not using out of spec RAM, then they should be using DDR4-3200 CL22. Anything else, is overclocked. <-- This is a fact. That is why computerbase did the test with DDR4-3,200 MHz, 22-22-22-42-2T, 1.2V (R9 3900X)

So basically everyone who is testing with DDR-3200 that is not CL22 is overclocking.

Which is pretty much everyone.

Which makes the whole DDR4-2933 thing for Intel quite hypocritical. Everyone's ok with OC the CL, but not the speed, eh?
 
Joined
Jun 13, 2012
Messages
1,232 (0.36/day)
System Name desktop
Processor i7-4770k
Motherboard Asus z87-plus
Cooling Corsair h80
Memory 32gb G.Skill Ares @ 2400mhz
Video Card(s) EVGA GeForce GTX 1080 SC (ACX 3.0)
Storage 2x Samsung 1tb 850evo SSD, 3x WD drives, 2 seagate
Display(s) Acer Predator XB271HU 27inch IPS G-Sync 165hz
Audio Device(s) Sound Blaster x-FI Platium, Turtle beach Elite pro 2 + superamp.
Power Supply OCZ Z Series 850W (10 years strong)
Mouse Logitech G502 hero
Keyboard Logitech G710+
Reading some these comments, amazing how dumb some people are. "Why don't you use xxxx memory" or "why don't you use RTX3090". There is A REASON for hardware they use when testing cause they have RESULTS using that same hardware to compare it to. The second you change memory or gpu you have nothing to compare it against cause every other test is using that hardware you just threw out. The results you get with different hardware now question's well was the gain cause the cpu is better or was it the faster memory you put in or was it the faster gpu you just dropped it.

Just as RandallFlagg said above https://www.techpowerup.com/forums/threads/amd-ryzen-9-5900x.274036/page-6#post-4386276
 

mkppo

New Member
Joined
Oct 30, 2020
Messages
8 (0.02/day)
AMD sent the 5950X so that it arrived today


After 3090, but not enough time to retest all comparison CPUs


Don't really care about what AMD recommends. I feel like 3200 CL14 is a great compromise between cost, performance and what's realistic for actual people who go out and buy hardware. I don't bench Intel with Power Limit removed, just because Intel wants it. Working on a memory scaling article for next week already

When doing the memory scaling article, kindly consider doing a test for 2x8GB vs 4x8GB as well. I believe it should marginally impact performance and would be very interesting to watch.

Awesome CPU launch. Amazing that there's 20%+ more performance at the same power draw on the same node. Kudos AMD.
 
Joined
Aug 24, 2004
Messages
201 (0.03/day)
I'm guessing there's not much to gain with regards to gaming with the 5950X? Was hoping for a review of that as well.
 

mkppo

New Member
Joined
Oct 30, 2020
Messages
8 (0.02/day)
But it is where the difference came from. People are talking about 0-5% difference in the reviews. Look at the 9900K at 2666 vs 3200, that's 14.4% to be precise. 3200->3600 is 1%. Those margins are more than enough for memory to explain this if AT ran their Intel chips at 2933 and AMD chips at 3200.

View attachment 174584

Many other websites tested at similar RAM speeds , including gamer nexus, and the 5900X seems to be slightly faster than 10900k.

Also, regarding this graph you repeatedly keep posting, keep in mind that Ryzen benefits massively from subtiming optimization as well. That, coupled with faster memory should increase FPS another 10%+ for Zen 3 as well. Have a look here (It's an older article for Zen 2):


It's generally known, and W1zzard himself mentions that Ryzen benefits more from faster RAM. But the main performance is obtained from subtiming optimization. My 3950x got atleast 10% more gaming performance at low res benchmarks, not that I care as I use it for rendering etc.
 
Joined
May 3, 2018
Messages
545 (0.41/day)
I'm guessing there's not much to gain with regards to gaming with the 5950X? Was hoping for a review of that as well.

Well Techspot tested the 5950X and it's the gaming champ for most part. They called it total domination!
 
Joined
Aug 24, 2004
Messages
201 (0.03/day)
Well Techspot tested the 5950X and it's the gaming champ for most part. They called it total domination!

Appreciate the link!

Edit: Uh, they used a top of the line new cpu and a 3090 and only benched games in 1080p!? Disappointing.
 
Last edited:
Joined
Dec 31, 2009
Messages
19,324 (4.43/day)
Benchmark Scores Faster than yours... I'd bet on it. :)
If you truly mean what you say about not using out of spec RAM

Which makes the whole DDR4-2933 thing for Intel quite hypocritical. Everyone's ok with OC the CL, but not the speed, eh?
I didn't say anything about that. The spec just lists the speed. JEDEC has to do with the memory itself. I'm talking about overclocking the IMC. It is not overclocking the RAM using the XMP profiles/what it's rated for on the box.

It isn't hypocritical at all. Now you're assigning a specific JEDEC specification from memory to something that doesn't have one. They just list speed, not specific timings. That is on the RAM itself. And again, if I buy a set of sticks rated to run DDR4 4000, they aren't overclocked. That is their specification. JEDEC profiles are on the sticks for compatibility reasons. Don't get your head twisted assigning meaning to things that aren't there. :)
 

W1zzard

Administrator
Staff member
Joined
May 14, 2004
Messages
23,297 (3.63/day)
Processor Core i7-8700K
Memory 32 GB
Video Card(s) RTX 3080
Display(s) 30" 2560x1600 + 19" 1280x1024
Software Windows 10 64-bit
Theree's Many things wrong with this review, really disappointing to see this from TPU.. TPU disappointment strikes yet again. I dont believe your 65c load temps on 10900k Its fake for sure.
Just gotta wait for the CPU to run into its power limit. For temps we're reporting the long term stable state, not the single highest temp. Does that make sense?
 
Top