AMD Ryzen 7 1800X CPU Review

[daddywalter]

@POMPOMPAIHN

So they made a chip that bests the Intel processors on some of the productivity tasks. Thats fine. But the majority of people are not buying this processor to run productivity tasks. They are mainly buying it for gaming. And to charge slightly more (at least in CDN $) than a 7770K and lose to it in all gaming benchmarks is horrible. I'd concede that these results would be fine if it was $100-$150 less than a 7770K. But this just seems like a mistake. When Intel releases 10nm processors probably early next year what is AMD going to do? It took AMD 4 years of development to basically get on park with 1ish year old Intel CPU's in gaming performance. They are in trouble when the 10nm Intel CPU's come out.

I'm sure AMD wanted Ryzen to be a great gaming chip that could beat Intel; but that didn't happen. I'm looking at its strengths and considering it for what it does best, which is more science- and business-oriented. Gamers drive technology, but business sales pay the bills at the high end of things. (More mainstream users don't need the most expensive products from either Intel or AMD; but they'll benefit as the new technologies become more affordable.)

The top Intel processors still beat Ryzen in some business/scientific tasks, while AMD wins in other benchmarks. I'm willing to call it a wash in general-purpose business computing, since I expect AM4 motherboards to cost less on average than Intel-based mobos, designed for similar purposes. Also, I expect both companies to lower prices for their processors now that Intel has serious competition again -- absolute-best performance will still command a premium, but processors lower in the companies' respective hierarchies should see price cuts that users will welcome.

Is Ryzen an undisputed Intel-killer? Of course not. But it makes AMD competitive with Intel to a degree the company hasn't seen in a long time. Competition is always good.

 

[Akai Miru]

Also keep in mind, this is with at launch drivers; so as time progresses so to will the fine tuning on BIOS, drivers and OCing with specific motherboards.

 

[ubercake]

I've looked at quite a few reviews on these today. I've also checked out a couple on the 1700x and 1700. The pricing for the R7s seems a little off. I think the 1700x and 1700 should be under the i7-7700K in price. Also, it seems like there are some things with the memory latency (mentioned in a few reviews) and limited PCIe lanes that make me think twice about making this my next platform. I'm curious to see how things pan out in the next couple of months.

 

[cabose369]

@daddywalter That is basically what I was kind of saying but you said it in more words. My beef is that it is too expensive for the performance you get. If the 1800X was $100-$150 less than a 7770K then it would be a great deal, not only just for productivity users, but also gamers, who want somewhat close to the performance of a 7770K, but for less. But instead, AMD tried to go for the high end, and came up short in gaming. Gaming is where you see people shelling out money (just wait for all the people that will buy up the 1080ti). To show so weak in gaming with your highest end CPU (considering it has double the cores and threads as the 7770K), and yet being more expensive, just doesn't make much sense.

 

[-Fran-]

Those complaining about the gaming benchmarks, please... The money is made in the SERVER realm, no the consumer (or prosumer) realm!

I am not justifying AMD having a bad showing there, but you guys need to understand priorities for a company that has to fight the behemoth Intel is.

You all have to put into context this: Intel has budget to do R&D for 2 BIG CPU families (3 if you count Atom) that cover very specific market segments. AMD has budget to do Zen and APUs based on the same CPU. Having the custom design as a side business based on the main CPU family.

The best take from this is: Zen has shown everyone around the world it is a solid foundation to continue developing CPUs based on this uArch and more than justifies the price for the prosumer world and will be a worthy adversary for Intel in the server realm. Something a lot of business will appreciate, I'm sure.

Cheers!

 

[Ts_1975]

Don't regret that I pre-ordered the 1700 and a b350 gigabyte motherboard (wish the board would arrive today). I decided that efficiency and value were more important than highest performance...it's also looking like the chip with the most headroom. The fact that so many motherboard makers are on board was my biggest reason to jump on board (if i recall correctly the original AMD Athlon had only 6 motherboards at launch in 1999).

I think the reviews and benchmarks are exactly what we should've expected. I know I'm not disappointed, and it doesn't change my opinion about Intel's best offerings. My hope is the ecosystem around the CPU continues to mature, to me it was one of the weakest points of an AMD platform. I certainly wasn't going to spend money on an AM3 or FM2 motherboard with DDR3 support.

This day was 10 years in the making, and I'm glad it finally arrived.

 

[randomizer]

Humbug. I'm sticking with Nehalem.

 

[FritzEiv]

Toms using the email Intel sent them to try and discredit AMD a little more. As AMD Suggested certain games, Intel suggests certain games. In previous CPU reviews, Tom uses other games but this one, not one game AMD suggested.

I grant you other websites show similar gaming fps reviews, but they also show other games where AMD is beating Intel and if its not beating it, its a draw.

We know who sponsors Tom's website.....

AMD suggested games came less than 24 hours before the embargo on this review lifted, along with a statement or two about gaming performance getting better in the future. Read into that whatever you'd like.

 

[boringpie]

Has anyone tested its performance on MS Office (ie. excel, access, powerpoint)? Also, other work application like SAS, Teradata.

 

[LilDog1291]

@Kiniku Wait until the Ryzen 3 and 5 series come out. Less cores with higher clock rates should translate into better gaming performance. When you are packing in 8 cores / 16 threads, heat demands that the clock rates be lowered or be faced with thermal throttling during gameplay anyway. Take a Ryzen 4 core/ 8 thread processor and compare it to the i5 or i7 series and see what the frame per dollar measurement is then.

 

[Giroro]

I don't understand the workstation results.
There are multiple places where the write up says 1800x falls behind, but the graph clearly shows it ahead (and vice versa).
For example single thread cinebench R15

"Using just one core at stock clocks, AMD’s Ryzen 7 1800X falls behind Intel’s Core i7-6900K a bit."

1800x = 158
6900k = 134
(Higher is better)
Something is wrong there, or am I missing something entirely?

 

[ibjeepr]

To add to my previous post after looking at many reviews online:

I think we need to wait to see what the price / performance of the 4 and 6 core Ryzens in comparison to the i7-7700k.
The R7 1800 series competes for business platform sales and does so well (even if it's not in the crushing win people were looking for) but definitely in the value aspect AMD promised. Most people here and online are looking for a gaming comparison when in reality this isn't about games.

If you a content creating gamer then this chip should be an exciting value, but if not, let's wait to see what performance / pricing the 4 and 6 core Ryzens offer when it comes to pure gaming. I'm hoping the 6 core is reworked just enough to bring 6 cores to hardcore gamers the way they just brought 8 cores to the business masses. If AMD leaves simultaneous multi-threading enabled, gets higher Ghz and does it at only 45 watts it could be very interesting. To be honest I'm not hopeful since the 4 and 6 core are said to compete with the i3 and i5 but I think there is a fair chance.

 

[matmat9v]

I think there is an error in Blender benchmark. I7-6900k default speed is 3.2Ghz and boosts to 3.7Ghz (or 4Ghz on Turbo Boost Max 3), so overclocking it to 3.8Ghz can at max give us a 20% advantage. How can it be explained that default 6900k does it's work in 53.27s and overclocked by 20% does the same in 34.19s - it is almost 36% faster.

 

[cordes85]

Dont let me down AMD i need an itx x300 for 1800x i want to build Antec razer cube, 2x 16gb green hand painted sticks, AMD Elite AM4 EKWB cpu block, i want the new NVIDIA gtx 1080ti, with ek gpu clear block with room for pump and reservoir, a 240mm slim rad, a slim 140 rad, green fittings, acrylic hardtube filled with Green Mayhems pastel colour, 1 x m.2 for Samsung Pro 2 Tb alone. Razer ripsaw linked to PS4 pro with HDD changed to 1 TB Samsung EVO Sata6, razer seiren pro, razer leviathian pro and with Creative Soundblaster X7 limited edition to complete the look, also using the Asus Rog Swift PG258Q 240hz 1080p monitor, green AK racing gaming chair. it all rests on whether AMD have the balls to what i need, preferably ASUS, Gigabyte, or MSI, but ill take anything and hand paint if needed the board

 

[FormatC]

I think there is an error in Blender benchmark. I7-6900k default speed is 3.2Ghz and boosts to 3.7Ghz (or 4Ghz on Turbo Boost Max 3), so overclocking it to 3.8Ghz can at max give us a 20% advantage. How can it be explained that default 6900k does it's work in 53.27s and overclocked by 20% does the same in 34.19s - it is almost 36% faster.

Turbo core works a bit tricky because you have on all cores totally different clocks. I tried that a few times and only in Blender gives you a fixed frequency (for this Intel) a big advantage. But I had not enough time to investigate more. The difference is less with smaller tile size.

But I also added the Blender Loop-Benchmarks. The picture is very different. I got the CPU only at Tuesday. No time to for more tests, sorry.

 

[n0ns3ns3]

I miss couple crucial test for me. in "workstation" realm there is nothing about Virtualization and code compilation performance 🙁

 

[akamateau]

As a consumer and an AMD enthusiast I would expect to find first run benchmarks that were run using Radeon dGPU cards.

Mantle, DX12 and Vulkan have eliminated the proprietary nVidia libraries that deliberately degrade AMD performance while running DX11.

SO the instead Tom's expects the reader to believe that running AMD's latest hot CPU with an nVidia dGPU will give any valid results?

ARE YOU FOLKS NUTS?

Enough of the DX11 benchmarks. Everybody and their dog knows that Microsoft no longer supports DX11 since 2015 and you continue to run DX11 as if it is relevant only so you can show Intel getting a leg up as DX11 API just does not run well on AMD hardware.

The you pair Ryzen with Nvidai dGPU which does NOT SUPPORT Asynchronous Compute Engines or Asynchronous Shader Pipelines. Then you have the unmitigated gall to declare that AMD Ryzen is sub-par.

WHY DON;T YOU USE A RADEON RX 480!!! HELL USE 2 AND CROSSFIRE MODE FOR dx11 AND EXPLICIT MULIT-ADAPTOR WILL KICK IN FOR dX12!!!

So before you lie to the consumer why don't you run hardware that Ryzen was designed and optimized for Radeon.

Or would that disagree with the ORDERS Intel gave you before you published this "review".

 

[akamateau]

Run the Ashes Benchmark again. This time use 2 Radeon RX 480.

2 x RX 480 is cheaper than 1 GTX 1080 AND crushes it with DX12 Explicit Mulit-adaptor and Asynch Compute. nVidia does not support Asynch Compute.

You might find the SMT issues go away with Radeon instead of the nVidai dGPU card.

Run Ashes on the Intel board with Radeon too.

You would if you had the guts.

 

[SpecialK27]

This is a $500 chip that from all the tests above, lags behind a $350 chip (i7 7700)
NOT ONLY in games, but in most desktop-office uses, and some workstation uses as well (like autocad, solidworks, etc.).
It's $150 more expensive than the i7 7700 and it only surpasses it, in rendering uses and handbrake. That is neither a win, nor what was suggested by amd before release. For all our sake i hope this can improve down the road with better support.

 

[jimmysmitty]

Toms using the email Intel sent them to try and discredit AMD a little more. As AMD Suggested certain games, Intel suggests certain games. In previous CPU reviews, Tom uses other games but this one, not one game AMD suggested.

I grant you other websites show similar gaming fps reviews, but they also show other games where AMD is beating Intel and if its not beating it, its a draw.

We know who sponsors Tom's website.....

You mean like the first game, AotS, that was a game that heavily favorited AMD before this review? And why use games suggested by AMD? Why use something that will only show a possible advantage and not pick what does perform better. That's the same marketing that AMD used for Bulldozer, where they would only show wins against the CPU it beat and not a fair overall view.

I think it is better to pick games that are more demanding and show a better idea of the performance itself, not what the manufacture recommends.

The increased height of the backplate threaded holes should make no difference to mounting force as long as the holes bottom out at the same depth as before. Unless the heatsink's screws have been designed to seat against the backplate's studs instead of bottoming out in the hole as the end-stop.

Most of the results were predictable: few games and mainstream applications scale beyond a few threads, so Ryzen gets beat in practically all of them. For most threaded compute-intensive stuff though, Ryzen does quite well in applications that haven't been extensively optimized for Intel.

I won't talk about any cooler specifically, but I test coolers. And the most-popular method for "stopping" the screw is to use an extended shoulder that bottoms against the top of the backplate's inset nuts.

So what did intel ask and/or tell you to do when they contacted Purch/Toms about the AMD Zen/Ryzen reviews?
Did you do as they asked or told them no like other sites have reported?

My last message from Intel was regarding the rescaling of a block diagram, on 28 December 2016. The response reads:
"Thanks for sharing. No issues from me."
There will be no Pigeon Holing Sessions today 😀

But...But Toms is an obvious Intel paid shill. They have never once said a good thing about AMD ever.

I don't get that. Do people want legit objective reviews or puff pieces?

Humbug. I'm sticking with Nehalem.

You silly goose you.

I think there is an error in Blender benchmark. I7-6900k default speed is 3.2Ghz and boosts to 3.7Ghz (or 4Ghz on Turbo Boost Max 3), so overclocking it to 3.8Ghz can at max give us a 20% advantage. How can it be explained that default 6900k does it's work in 53.27s and overclocked by 20% does the same in 34.19s - it is almost 36% faster.

If it is on all cores the i7 will not clock to its max turbo, however when OCing it normally clocks it to that as a max and not just a turbo speed.

 

[g-unit1111]

I knew Ryzen was not going to be perfect from the beginning but from the reviews that I have seen start to trickle out it's a huge step in the right direction for AMD. Going from a 220W mess of a CPU to a 95W one is not just the right move, it will be a gigantic step forward. Having one company control an entire industry like that is not a good thing. Any competition for Intel is good for the industry and drives progress. Ryzen is the first release from AMD that I have been excited for in quite some time. And it has definitely met my expectations. Some of the AM4 motherboards look amazing, especially that MSI one. Well done AMD.

 

[mapesdhs]

Typos in the Python NumPy diagram(s), the annotations are all mixed up, titles for the Ryzen CPUs don't match the colours (swapped with what are presumably supposed to be 7700K entries). Another typo in the CB R15 multithreaded image, says "XFR Endabled" (doesn't someone read this stuff before commiting changes?).

Also, on page 4, the "Interface" entry for the 8350 states the process size instead of the socket name.

For the Blender "Ryzen" workload, this comment, "AMD’s Ryzen 7 1800X beats Intel’s Core i7-6900K at the same clock frequency, but falls behind at stock clock frequencies.", is the complete opposite of what the graph shows.

Typo in the titles for the large composite scores of 3ds Max 2015 ("composit").


Btw, the popup video ad things are a complete and total pain. They keep getting in the way of the navigation menus. Please fix them or preferably just get rid of them, they're really annoying.

Ian.

 

[nader21007]

So if Intel cant beat AMD 1800X with the $1000 6900K, no problem, it could use "Call us before you write $$$" , oh dear Tom use sth like 3.8 Ghz gimmick and pant Ryzen bad. Well in gaming put Ryzen againt the 7700K, when testing for Multi-threaded tasks , put Ryzen against 6900k , ....
do you know what is the 3.8 Ghz?
If we compare only Ryzen 1800 with 6900K, Ryzen will be a winner, albait without the ingridients that Tom added for $$$$.( 3.8 Ghz & 7700K)

 

[BNCID]

As a bias and an devout AMD fanboy. How dare you say anything other than AMD is great and everything else sucks and anybody who says differently is a paid shill

Fixed and summarized that for yah.

 

[jimmysmitty]

So if Intel cant beat AMD 1800X with the $1000 6900K, no problem, it could use "Call us before you write $$$" , oh dear Tom use sth like 3.8 Ghz gimmick and pant Ryzen bad. Well in gaming put Ryzen againt the 7700K, when testing for Multi-threaded tasks , put Ryzen against 6900k , ....
do you know what is the 3.8 Ghz?
If we compare only Ryzen 1800 with 6900K, Ryzen will be a winner, albait without the ingridients that Tom added for $$$$.( 3.8 Ghz & 7700K)

Are you not looking at the same review as I am? Because every single benchmark has the same CPUs in it, mainly the best PD, a 7700K, a 6900K and the 1800X. Gaming , productivity and workstation are all testing the same CPUs. That is a fair comparison.

And the 3.8GHz is an advantage for Ryzen against the 7700K because the 7700K is stock clocked at 4.2GHz base and 4.5GHz turbo and the main competition for Ryzen will be the 7700K, not the 6900K as the 6900K is a niche market. Hell even the 6800K is a niche market CPU as the majority of people buy either a 7600K or 7700K. Either way it puts all CPUs on a level field clock speed and allows us to see what the performance of each are from a clock speed point of view. This has been done in plenty of reviews to give a baseline.

And seriously, you really think Intel pays Toms? I can tell you they do not. Intel doesn't need to as for the past few years AMD has been almost non-existent in the competitive market.