***OFFICIAL*** Ryzen 5000 / Zen 3 Launch Thread REVIEWS BEGIN PAGE 39

[uzzi38]

339-284 = $55 increase over the span of almost a decade. I'm not even sure if it accounts for the inflation for the same period.

See this whole argument is stupid because you can just pick out arbitrary dates like you are.

I could turn around and look at the jump from the 860 to the 5775C for example. It's meaningless.

Or I could turn around and compare the 5800X to the 1800X. Would be pretty similar to what you're doing with the 5800X now, describing a price hike without knowing/ignoring the rest of the SKUs that may or may not come.

Ultimately I posted to prove you wrong on the statement that Intel did not take raise prices for the same core/thread count from one generation to the next. Anything past that is a ridiculous argument because of the fact we don't know how pricing will fluctuate for a generation of product that will last another year and a few months.

 

[Makaveli]

Also 5900X runs quite cooler than 5800X:

https://tpucdn.com/review/amd-ryzen-9-5900x/images/cpu-temperature.png

https://tpucdn.com/review/amd-ryzen-7-5800x/images/cpu-temperature.png

So, if I had the money I'd buy the 5900X in an instant however given the new pricing policy I'm not touching Ryzen 5000 until prices go down and there's a sale. Besides I'm waiting for 5700X regardless - I've no desire to put a 142W CPU in my case.

This review doesn't show that. I've seen some other inconsistency with TPU numbers like veracrypt being low. I think they need to re do there test bed.

AMD Ryzen 7 5800X review

It's time for already our 4th ZEN3 review, yes the much anticipated Ryzen 5 5800X. This is the processor that is on the watchlist of many with 8 cores and 16 threads if offers a bit more flexibility ... Power Consumption and temperatures

www.guru3d.com

 

[jpiniero]

Ultimately I posted to prove you wrong on the statement that Intel did not take raise prices for the same core/thread count from one generation to the next. Anything past that is a ridiculous argument because of the fact we don't know how pricing will fluctuate for a generation of product that will last another year and a few months.

I would not expect prices to fall, or there to be cheaper SKUs available to DIY. Unless Rocket Lake delivers AND is available... neither of which is a given.

The difference with Zen 3 is that with the previous gen products AMD was constrained on pricing due to gaming perf.

 

[yeshua]

This review doesn't show that. I've seen some other inconsistency with TPU numbers like veracrypt being low. I think they need to re do there test bed.

AMD Ryzen 7 5800X review

It's time for already our 4th ZEN3 review, yes the much anticipated Ryzen 5 5800X. This is the processor that is on the watchlist of many with 8 cores and 16 threads if offers a bit more flexibility ... Power Consumption and temperatures

www.guru3d.com

W1zzard is quite open about their testing methodology, so I have no reasons not to trust him. According to the Guru3D reviews (which used a different cooler, had a different case and most likely a different environment temperature):

5800X under load: 73C
5900X under load: 70C

So, again the 5900X runs cooler, so my initial statement holds true. It can be easily explained by the fact that the 5900X features two six-core CCX'es and 5800X just a single one, so the heat density of the 5800X is going to be higher.

 

[uzzi38]

I would not expect prices to fall, or there to be cheaper SKUs available to DIY. Unless Rocket Lake delivers AND is available... neither of which is a given.

The difference with Zen 3 is that with the previous gen products AMD was constrained on pricing due to gaming perf.

Well it also depends on how popular Milan is and how much demand ramps up over the next few months as well.

As you can probably imagine, it is really quite high right now (and is the main reason why the 5800X is the worst value product - the 64 core Milan SKUs are the most popular and the 5800X is much lower margin than anything Milan and also anything 5950X).

 

[Makaveli]

W1zzard is quite open about their testing methodology, so I have no reasons not to trust him. According to the Guru3D reviews (which used a different cooler, had a different case and most likely a different environment temperature):

5800X under load: 73C
5900X under load: 70C

So, again the 5900X runs cooler, so my initial statement holds true. It can be easily explained by the fact that the 5900X features two six-core CCX'es and 5800X just a single one, so the heat density of the 5800X is going to be higher.

It runs cooler yes but your first example was a 6c spread and this one is now 3c. You are correct that heat density will be slightly higher on the 5800X but it won't be enough to really matter.

 

[exquisitechar]

What a win against a five years old uArch running on a much inferior node! Congrats! AMD fans should celebrate it until the rest of their lives. What next, should we compare Ryzen 5000 to Sandy Bridge CPUs for good measure? Alright! An amazing win!

You can compare it to Tiger Lake and Zen 3 will still mostly be ahead.

 

[jpiniero]

Well it also depends on how popular Milan is and how much demand ramps up over the next few months as well.

I kinda forgot about Milan, esp since they haven't officially announced it yet. That only makes it less likely. OEMs would prefer lower TDP models but there is zero reason to also make them available to DIY.

 

[yeshua]

You can compare it to Tiger Lake and Zen 3 will still mostly be ahead.

It's all quite confusing to me and something about TGL does not feel right as its performance has regressed vs ICL in certain areas, so I don't want to pay much attention to Intel at the moment because their product line has been an unmitigated mess for a couple of years now. What I want to see a comparison of TGL vs Ryzen in pure ST workloads at a frequency where TGL won't heavily throttle, e.g. 3GHz at most and then make conclusions based off that. Intel's own 10nm node is seemingly unfit for their uArch. Anyways it's all kinda off-topic here as we are discussing the new Ryzens.

It runs cooler yes but your first example was a 6c spread and this one is now 3c. You are correct that heat density will be slightly higher on the 5800X but it won't be enough to really matter.

Again, different coolers, thermal pastes and their application, environments and I'm sure different workloads. Almost all reviews have different maximum temperatures of new Ryzen CPUs and all I was trying to say is that 5900X runs cooler. Isn't that the case? ;-)

 

[Dave3000]

W1zzard is quite open about their testing methodology, so I have no reasons not to trust him. According to the Guru3D reviews (which used a different cooler, had a different case and most likely a different environment temperature):

5800X under load: 73C
5900X under load: 70C

So, again the 5900X runs cooler, so my initial statement holds true. It can be easily explained by the fact that the 5900X features two six-core CCX'es and 5800X just a single one, so the heat density of the 5800X is going to be higher.

Now I wonder if I should go for the 5900x instead of the 5800x for just $100 more. I have a Noctua U12A cooler. Does anyone here know if the starting throttle temperature is higher on the 5800x than the 5900x?

 

[shady28]

W1zzard is quite open about their testing methodology, so I have no reasons not to trust him. According to the Guru3D reviews (which used a different cooler, had a different case and most likely a different environment temperature):

5800X under load: 73C
5900X under load: 70C

So, again the 5900X runs cooler, so my initial statement holds true. It can be easily explained by the fact that the 5900X features two six-core CCX'es and 5800X just a single one, so the heat density of the 5800X is going to be higher.

Guru3d used DDR4-3600 on all platforms, while TPU uses DDR4-3200. Zen 3 doesn't seem to really come into its own unless it is paired with the DDR4-3600. AnandTech by comparison, used DDR4-2933 on Intel and DDR4-3200 on AMD. Personally I prefer the test platforms Guru3d and TPU use because it is more reflective of an enthusiast build, while AnandTechs is more reflective of say a Dell XPS Intel vs a custom build AMD (which is meaningless). But to really get a full understanding you have to look at all the different sites benchmarks with an eye on the test configurations. That really shows that the winning setup is a Zen 3 with DDR4-3600. Get lower than that and from what I can tell, you might as well have an Intel.

 

[lobz]

Read the user's name backwards(<-) and you will figure it out

Can someone help me unread this?!!!

 

[Dave3000]

Guru3d used DDR4-3600 on all platforms, while TPU uses DDR4-3200. Zen 3 doesn't seem to really come into its own unless it is paired with the DDR4-3600. AnandTech by comparison, used DDR4-2933 on Intel and DDR4-3200 on AMD. Personally I prefer the test platforms Guru3d and TPU use because it is more reflective of an enthusiast build, while AnandTechs is more reflective of say a Dell XPS Intel vs a custom build AMD (which is meaningless). But to really get a full understanding you have to look at all the different sites benchmarks with an eye on the test configurations. That really shows that the winning setup is a Zen 3 with DDR4-3600. Get lower than that and from what I can tell, you might as well have an Intel.

Well, the 10700k officially supports up to DDR4-2933 and the Ryzen 5000 series support officially up to DDR4-3200 and if someone just wanted to build a system to conform to official specifications, that person would be most likely running DDR4-2933 if he got a 10700k or DDR4-3200 if he got a 5800x, 5900x, or 5600x. So even within officially supported specifications, that person would still be running faster RAM (3200 instead of 2933) if he got one of those Ryzen processors instead of a 10700k. I'm not talking about enthusiats or overclockers here.

 

[lobz]

A nice chart: shows just how great the Sky Lake uArch is as it's been able to compete successfully for five years in a row with almost zero underlying changes (aside from the number of cores/frequencies and microarchitectural tweaks to fix HW vulnerabilities like Meltdown). Thanks! Also puts to rest the claims that the Ryzen 5000 series is universally faster than ages old Sky Lake.



More than.



I don't know a single person who runs power viruses 7x24.



It's because both 5600X and 5800X are overpriced junk in terms of performance per dollar - and this is where AMD has lost their mind. Many-cores CPUs should be a lot more expensive per core because they are used by professionals who actually can recuperate their cost.

This time around both 5600X and 5800X are absolute worst in this regard.



What a win against a five years old uArch running on a much inferior node! Congrats! AMD fans should celebrate it until the rest of their lives. What next, should we compare Ryzen 5000 to Sandy Bridge CPUs for good measure? Alright! An amazing win!




It must take a lot of effort to find tasks which are going to be seriously affected by running on two CCX'es simultaneously. 5800X defeats 5900X in certain scenarios but by a very small margin, however if you're going to run truly parallelized tasks like rendering, video encoding, compilation, AI - then obviously 5900X will be a much better choice not to mention it costs less than 5800x in terms of price per core.

Also 5900X runs quite cooler than 5800X:

https://tpucdn.com/review/amd-ryzen-9-5900x/images/cpu-temperature.png

https://tpucdn.com/review/amd-ryzen-7-5800x/images/cpu-temperature.png

So, if I had the money I'd buy the 5900X in an instant however given the new pricing policy I'm not touching Ryzen 5000 until prices go down and there's a sale. Besides I'm waiting for 5700X regardless - I've no desire to put a 142W CPU in my case.

If you call the 5600X overpriced junk, what have you been calling the 10600K? I swear I don't wanna search for your comments, so I'm sorry if you already stated your opinion before. I'm really curious.

 

[lobz]

W1zzard is quite open about their testing methodology, so I have no reasons not to trust him. According to the Guru3D reviews (which used a different cooler, had a different case and most likely a different environment temperature):

5800X under load: 73C
5900X under load: 70C

So, again the 5900X runs cooler, so my initial statement holds true. It can be easily explained by the fact that the 5900X features two six-core CCX'es and 5800X just a single one, so the heat density of the 5800X is going to be higher.

TPU has been also quite open about their Ryzen 5000 tests results being reeeeeally strange. Kudos to them, actually.

 

[shady28]

Well, the 10700k officially supports up to DDR4-2933 and the Ryzen 5000 series support officially up to DDR4-3200 and if someone just wanted to build a system to conform to official specifications, that person would be most likely running DDR4-2933 if he got a 10700k or DDR4-3200 if he got a 5800x, 5900x, or 5600x. So even within officially supported specifications, that person would will be running faster RAM if he got one of those Ryzen processors instead of a 10700k.

Ya, I just had this conversation with someone at Toms. The official spec is using JEDEC standards. Guess what JEDEC standards are for CL? CL 19 for DDR4-2933, and CL20 for DDR4-3200. So basically they are not using the standards to start with, and overclocking CL is overclocking. It's also ludicrous to state that someone with a 10700K is going to use DDR4-2933. That would only happen on OEM systems, which tend to adhere to the JEDEC standard.

Hence my analogy of Anands tests being a simulation of a Dell XPS vs Custom build. I think that is a comparison, but how useful is it on a site visited mostly by custom builders. In the real world, the Z490 overclocks RAM way better than anything I've seen on the AMD front - and it shows on places like PCMark where Intel rigs with DDR4-4800 dominate.

 

[lobz]

Ya, I just had this conversation with someone at Toms. The official spec is using JEDEC standards. Guess what JEDEC standards are for CL? CL 19 for DDR4-2933, and CL20 for DDR4-3200. So basically they are not using the standards to start with, and overclocking CL is overclocking. It's also ludicrous to state that someone with a 10700K is going to use DDR4-2933. That would only happen on OEM systems, which tend to adhere to the JEDEC standard.

Hence my analogy of Anands tests being a simulation of a Dell XPS vs Custom build. I think that is a comparison, but how useful is it on a site visited mostly by custom builders. In the real world, the Z490 overclocks RAM way better than anything I've seen on the AMD front - and it shows on places like PCMark where Intel rigs with DDR4-4800 dominate.

So what? You can almost buy a 5900X for the price of 16 gigs of DDR4 4800. The current limit on Zen 3 seems to be between 3900 and 4000 on average, with very good samples going beyond that. So instead, show me an instance where, for example, an intel system with 4200+ RAM dominates a Zen 3 system with 3933-4000 RAM.

I'm not 'challenging' you per se, I'm just saying, it is in fact easier to go really overboard with RAM speed on intel, I'm just not sure what realistic sense it makes over 4000 MHz, given how fast the RAM price gets out of control. It's been a long journey for Zen to get there, but it's looking like a real all-rounder this time. Intel and AMD switched sides and it's the blue team who seems to be trying to hold on to very-very narrow niches.

edit: of course I'm assuming 1:1 IF on AMD and controlled timings on both, people throwing money into fast RAM usually try to do it right.

 

[shady28]

So what? You can almost buy a 5900X for the price of 16 gigs of DDR4 4800. The current limit on Zen 3 seems to be between 3900 and 4000 on average, with very good samples going beyond that. So instead, show me an instance where, for example, an intel system with 4200+ RAM dominates a Zen 3 system with 3933-4000 RAM.

Well, you just want hyperbolic. All I've really said here is that you need DDR4-3600 to get a solid win over Intel. And while DDR4-4800 is expensive, it's not as expensive as you state. Step it down just a bit to DDR4-4400, which AMD can't run, and you're looking at $125 for 16GB while DDR4-4600 is $199.

TPU is supposed to do a memory scaling comparison with Zen 3, hopefully will include some Gen 10 parts. That should be interesting and resolve the question as well.

 

[lobz]

Well, you just want hyperbolic.

Nooo.. no no no I don't. I also edited the post, anticipating you'd get the wrong idea. Please re-read it and answer accordingly, thank you and sorry.

TPU is supposed to do a memory scaling comparison with Zen 3, hopefully will include some Gen 10 parts. That should be interesting and resolve the question as well.

A lot of other sites will do that as well and TPU will be one of those I probably won't watch. The only thing they're doing worse than CPU benching is GPU benching. That's my opinion after a decade of trying to figure out how the fudge they get results that are almost always differ from the aggregate of almost all other trustworthy sources. They're not leaning towards any direction, in my eyes their results are just unusable.

 

[mikk]

The gaming performance looks underwhelming to me. It's a bit faster than CML-S with a RAM speed advantage. With the same RAM the gaming peformance seems to be on par:

AMD Ryzen 9 5900X Review

The Ryzen 9 5900X dominates Intel's Core i9-10900K in our testing because of AMD's massive IPC improvements. At $550, this processor is certainly not cheap, but it offers so much more performance, especially single-threaded, that AMD has a clear winner on their hands.

www.techpowerup.com

Considering that CML-S is based on the Skylake architecture this isn't great.

 

[lobz]

The gaming performance looks underwhelming to me. It's a bit faster than CML-S with a RAM speed advantage. With the same RAM the gaming peformance seems to be on par:

AMD Ryzen 9 5900X Review

The Ryzen 9 5900X dominates Intel's Core i9-10900K in our testing because of AMD's massive IPC improvements. At $550, this processor is certainly not cheap, but it offers so much more performance, especially single-threaded, that AMD has a clear winner on their hands.

www.techpowerup.com

Considering that CML-S is based on the Skylake architecture this isn't great.

I guess you haven't read a single message in this thread. Look for better sources than TPU, that's my advice.

 

[Dave3000]

Well, you just want hyperbolic. All I've really said here is that you need DDR4-3600 to get a solid win over Intel. And while DDR4-4800 is expensive, it's not as expensive as you state. Step it down just a bit to DDR4-4400, which AMD can't run, and you're looking at $125 for 16GB while DDR4-4600 is $199.

TPU is supposed to do a memory scaling comparison with Zen 3, hopefully will include some Gen 10 parts. That should be interesting and resolve the question as well.

I'd like to see a review where a 10700k is running JEDEC DDR4-2933 comparing to a 5800x running JEDEC DDR4-3200. Not everyone wants to assemble a computer and would rather buy a system from Dell, HP, etc and would most likely choose a system based on their choice of CPU primarily. These systems don't use XMP memory.

 

[shady28]

Nooo.. no no no I don't. I also edited the post, anticipating you'd get the wrong idea. Please re-read it and answer accordingly, thank you and sorry.

A lot of other sites will do that as well and TPU will be one of those I probably won't watch. The only thing they're doing worse than CPU benching is GPU benching. That's my opinion after a decade of trying to figure out how the fudge they get results that are almost always differ from the aggregate of almost all other trustworthy sources. They're not leaning towards any direction, in my eyes their results are just unusable.

No worries. You can find a lot of examples on PCMark site right now. Pull up time spy, single GPU, dual channel memory (to eliminate HEDT systems), sort by CPU score, limit to 5900X and then 5950X. The highest scoring 5950X and 5900X are running at DDR4-3800. Do the same to the Intel 10900K and set max core speed to 4900 (the max all core turbo Intel specs on the 10900K). The Intel rigs in this range are running DDR4-4400 to DDR4-4600.

 

[shady28]

What about businesses and wo


I'd like to see a review where a 10700k is running JEDEC DDR4-2933 comparing to a 5800x running JEDEC DDR4-3200. Not everyone wants to assemble a computer and would rather buy a system from Dell, HP, etc and would most likely choose a system based on their choice of CPU primarily. These systems don't use XMP memory.

I agree. I wish they would do some reviews of OEM rigs, or custom builds with OEM rig type specs. Most don't look that far in depth though and I'm sure they'd get flack for it.

 

[Veradun]

What about businesses and wo


I'd like to see a review where a 10700k is running JEDEC DDR4-2933 comparing to a 5800x running JEDEC DDR4-3200. Not everyone wants to assemble a computer and would rather buy a system from Dell, HP, etc and would most likely choose a system based on their choice of CPU primarily. These systems don't use XMP memory.

Isn't Ian using exactly those specs in his CPU reviews?