AMD Ryzen 7 1800X 3.6 GHz

[G33k2Fr34k]

It would've been nice to see a 6900K along side the 1800X just for reference. The comparison against the 7700K is certainly good to see but the 7700K is not 1800X's competitor. The power consumption figures are amazing!

I'm waiting for the 1400X 3.9GHz quad core CPU. I'd love to see a review of it compared against the 7700K when it comes out.

 

[Lightofhonor]

Went with a 1700X since, when I preordered at least, it wasn't completely known if the 1700 would be able to achieve similar overclocks.

Overall great review! I agree the MB issues are a pain, but even with those I am still way outperforming my old i5 with 1333mhz memory And it will just get faster.

I do wish we could keep XFR/Boost while overclocking. Enjoying my 3.8ghz, but knowing I am actually slower in single core makes me want to OC more...

 

[mcraygsx]

This is what annoys the crap out of me(Intel is guilty too). Why change the mounting hole layout so we have to buy(or the heatsink manufacturers have to give away at a loss) new retention brackets? Did Intel really need to make the holes on the 115X platform ever so slightly larger than the 775? Did AMD really need to do the same between AM3+/FM2+ and AM4? And it makes even less sense that AMD wouldn't take the opportunity to make their mounting holes square, and they should have actually just matched the already in use Intel spacing. Make it easier for all of us!

Asus Crosshair VI support both AM3/AM4 since it has 8 holes total for cooler. You can upgrade to new AM4 cooler or extend the life of your existing AM3 Cooler/Kit. Motherboard manufacturers can also make this transition easier should they choose to do so.

Therefore I wouldn't put much blame on AMD. Since Intel does it each season.

 

[Ubersonic]

Not all that impressive for gaming, you're still better off with a i7 7700K for gaming.

To be brutally honest there is zero reason for anyone with a brain to buy a 7700K now. If a Ryzen chip isn't better than a 7700K for what you're doing, then what you're doing doesn't warrant buying the 7700K over the 7600K either.

 

[newtekie1]

Asus Crosshair VI support both AM3/AM4 since it has 8 holes total for heatsink.

Yeah, there were some boards that did that with the 775/1156 transition too. That just shows even more proof that there was no reason to change the hole layout. It is nice that some manufacturers recognize this, but it is something that AMD and Intel should recognize.

To be brutally honest there is zero reason for anyone with a brain to buy a 7700K now. If a Ryzen chip isn't better than a 7700K for what you're doing, then what you're doing doesn't warrant buying the 7700K over the 7600K either.

Agreed. If the extra threads of the 7700K are going to help you, then the extra cores of the 1700 will help you even more. And the gaming performance of the 1700(with an overclock) will be good enough.

 

[XiGMAKiD]

Excellent review, the power consumption is impressive. Now if only AMD can use that power consumption advantage to make it perform and overclock better

Just curious, with Ryzen 5 already announced and launching worldwide on April 11, do TPU already got one?

 

[Shatun_Bear]

It's funny how many times hardware reviewers get accused of being anti-AMD, while they are doing their damnedest to paint its products favorably as if their livelihood depended on it, because it does.

So then why is this being compared against a 7700K and nothing else? Last I checked TPU wasn't a 'gaming' website but a 'hardware' website. AMD didn't compare it to a 7700K for a reason. It is a HEDT competitor - so at least put the 6900K, 6950K or a 6800K in there.

DX12 / Vulkan


It's a first-round sample. I just didn't have time.

But you said this:

Arguably the most important product launch for AMD as a processor company

So with a review over 2 weeks after launch, I take it AMD is pretty low down in your estimations?

Yet the Nvidia 1080 Ti gets reviewed a day before it launches.

 

[Vario]

Hardly.

The low score was predictable after AMD stopped sending TPU first-round review samples.

It seems anything Nvida = 9.7-9.9 score, anything AMD as low as 8.7. Really transparent bias.

As far as gaming goes, put simply Ryzen is more money and performs worse and with more hassle. Its scored to account for that.

 

[dozenfury]

On the good side AMD is closer than they've been in many years. But on the downside unless you do not game and primarily do heavy productivity/encoding (and are counting every second at that), the 7700K will be the better option. It's much cheaper than the 1800X and essentially the same price as the 1700X. Even then the 7700K can still do the productivity apps with fine performance, and is far more overclockable which would increase the gap even further than the results in this review. I was hopeful for Ryzen, and I don't have a bias either way since I've had plenty of Intel and AMD systems. But it's tough to argue for a 1700/X or 1800X as-is. This is a bit harsh, but realistically with it's performance, lack of oc headroom, and various minor outstanding issues, the 1800X would need to be in the mid $200's to be a buy for me over a $320 7700K.

Not to mention if I'm encoding something, I'll start it and go do something else for awhile while it runs or just let it run in the background. So the amount of time that takes for passive activity like encoding is really far less important than gaming or general use response time, where you are actively interacting and are more directly affected by the performance difference. FPS and response time is important. Faster encoding time on the other hand is a bit like saying my oven can cook a pizza in 18 minutes instead of 20. Nice, but something that's tough to get super excited about.

 

[the54thvoid]

Hardly.

The low score was predictable after AMD stopped sending TPU first-round review samples.

It seems anything Nvida = 9.7-9.9 score, anything AMD as low as 8.7. Really transparent bias.

As someone who has just purchased an entire new Ryzen system, I feel qualified to reply to this.

I've gambled. I've bought 3200 memory that is on a QVL list. It should just work. It better. I've just bought a £370 CPU that will game worse than a 7700k. But I'm hopeful some optimising will be done to address the CCX latency (thus the 3200 RAM speed). I know it wont clock much over 4Ghz, (if i'm lucky). I actually have no idea why i bought a Ryzen CPU. All I do is game. I just wanted a change and you know what. If it was a mistake and I have issues, I'll come back and ram it down your throat that the Ryzen path is a perilous one.

You know why Nvidia cards often get 9 point whatever? Because they are very good. You did notice AMD used a Titan Xp for it's open Ryzen demos? And last outing for Fury X type PC system that never really happened they used an Intel chip. Join the dots and don't be so AMD banner waving blind.

In fact - I think I bought an AMD chip so I can get all pious about it and tell the fanboys to go jump off a bridge.

 

[newtekie1]

So with a review over 2 weeks after launch, I take it AMD is pretty low down in your estimations?

Yet the Nvidia 1080 Ti gets reviewed a day before it launches.

I'm sure the 1080 Ti was a lot quicker to review than Ryzen was. The test bed was already set up for the 1080 Ti review. It was just a matter of popping the card in, taking a few pictures, copy and pasting some information into the review template and done.(Sorry, W1zzard. I know this is a simplification of all the work you put into the reviews. I'm just trying to make the point that you have the process down to a science for GPU reviews. So they are a lot quicker to do and get published.)

Testing a brand new CPU, when you aren't using to testing CPUs, is a little different.

Plus, W1zzard isn't exactly a CPU reviewer. Most of the CPU reviews have been done by others. So I can see it taking him a little longer to get the review done and get the process down. Plus, he is extremely thorough, which is one of the reasons I really like his reviews. And he is dealing with a brand new platform that several reviewers, some way more expereience with CPU reviews, even had problems dealing with. Hell, just dialing in a stable overclock can take a week of fiddling and stress testing...

 

[Big_Vulture]

It took quite really long to TPU...

 

[efikkan]

For anyone planning to game on a GTX 1070 or better, this CPU is simply not good enough. An i7-6800K would be a better buy then, even with "only" 6 cores.

 

[ZeroFM]

Only problem in test is : all games is AAA , and it's best case scenario for amd . In arma 3 1800x vs 7700K is 30%+ FPS different , same should be in ARK , Rust , The forest

 

[xkm1948]

Cons: Lacks integrated graphics?

Is W1zzard smoking something? Who the f*uck in the right mind would want an iGPU on an 8 core HEDT processor? It is like anyone actually uses Intel's shit iGPU when they buy 6700K/7700K.

 

[W1zzard]

I'm sure the 1080 Ti was a lot quicker to review than Ryzen was. The test bed was already set up for the 1080 Ti review. It was just a matter of popping the card in, taking a few pictures, copy and pasting some information into the review template and done.(Sorry, W1zzard. I know this is a simplification of all the work you put into the reviews. I'm just trying to make the point that you have the process down to a science for GPU reviews. So they are a lot quicker to do and get published.)

Testing a brand new CPU, when you aren't using to testing CPUs, is a little different.

Plus, W1zzard isn't exactly a CPU reviewer. Most of the CPU reviews have been done by others. So I can see it taking him a little longer to get the review done and get the process down. Plus, he is extremely thorough, which is one of the reasons I really like his reviews. And he is dealing with a brand new platform that several reviewers, some way more expereience with CPU reviews, even had problems dealing with. Hell, just dialing in a stable overclock can take a week of fiddling and stress testing...

Right on the money, this is my first CPU review. Which means selecting and figuring out benchmarks, then building test systems with the hardware that's available, then bench (not exactly few results), then think, fix bench suite, rebench everything (two times for this review), then come up with structure, layout, texts, conclusion.

There will be more CPU reviews from me though Just bought i5 7400, i3 7100, Pentium G4560.

 

[the54thvoid]

I'm going to crank AA to 64x and put tesselation to warp factor. That'll reduce my CPU bottleneck.

 

[xkm1948]

If it was me testing, I would probably push it through a whole suite of Linux based super computing performance test as well as Bioinformatics test. At least you need to use some BWE/HWE for comparison!

 

[W1zzard]

as well as Bioinformatics test.

Do you have a workload you could share? Which is not running much faster on GPU

 

[xkm1948]

Do you have a workload you could share? Which is not running much faster on GPU

A lot of genomic stuff I do cannot be done on GPU due to lack of VRAM. Protein 3D folding on the other hand relies heavily on GPU(Standford Fold@Home)

http://www.genome.umd.edu/masurca.html
This is what I use for genome assembly. CPU performance matters a lot in these computation intensive linux based testings. MaSuRCA is totally free. I think most RAW human genome reads are also free for grab on NCBI as sequencing reading archives. I used to use 30X human genome reads as performance test for PCs. Assembly of a 30X coverage human genome takes about 166hrs on an 8 core 16 thread IvyBridgeE Xeon. The same amount of work it took 5~6hrs on my overclocked 5820K. On a HWEP based 20 core 40 thread super computing node it takes ~3hrs.

And this is de novo(without reference) assembly I am talking about.
So yes for us Biofimaticians RAW CPU performance is very important.

Example of one human genome sequence reading archive(SRA)
https://www.ncbi.nlm.nih.gov/sra/ERX1943173[accn]

 

[Air]

I think ryzen will shake up things in gaming notebooks, where power consumption is crucial and framerates higher than 60 fps are rarely the target.

 

[londiste]

in anandtech's ryzen strictly technical thread (https://forums.anandtech.com/threads/ryzen-strictly-technical.2500572/) stilt wrote:

Note: Current versions of Prime95 (28.10) do not stress Ryzen CPUs properly. The resulting power consumption is abnormally low, and both Firestarter and Linpack result in significantly higher power consumption.

is there any truth to this?

 

[1c3d0g]

@W1zzard: no distributed computing benchmarks? I'd love to see if AMD's 2x cores beats Intel's higher clocks with SETI@Home (or some other BOINC project) work units.

 

[the54thvoid]

To be fair, on a gaming front, having looked at the 1440p results, it's good news for me.

 

[CJCerny]

If most of your computing time is spent gaming, sure seems like the 7700K is still the way to go, especially when you consider that it is at least $100 cheaper than this Ryzen CPU at your local Microcenter. If you spend most of your time using multi-core liking apps, then this is a product to consider. I'm not seeing any bias in this review--buy what suits your workload best. You don't need a $500 CPU to play games well and surf the web.