AMD Ryzen 1700X CPU Review @ [H]

[Simplyfun]

It is design choice, though tech report review really made me reconsider how long this gen of Ryzen will last in workstation tasks.

I don't really see this Ryzen as a workstation CPU at all actually. I mean , the limited RAM support alone... what the hell am I going to do with 64Gigs? That will be Naples.

I mean I was thinking it over, a 1700, maybe disable the turbo and keep it cool and calm. 16 lanes available means some compromises in RAID controller choices and Compute GPU. Sure you can just settle for 8 lanes for RAID and 8 for GPU and leave a couple gen 2.0's for networking. But then ... 64 Gigs... pointless for me anyhow. Plus the boards coming out are all home oriented so far.

Too many compromises for me, but I'm sure it will fit the niche for some.

Anyhow, based on what I've seen, I like the cores a lot. I like how everything works and I think a Naples workstation might be priced right when they get here. In the meantime I'll keep on keeping on with my Xeons. There's no rush.

 

[gigaxtreme1]

It is design choice, though tech report review really made me reconsider how long this gen of Ryzen will last in workstation tasks.

AVX-512 handling can be written into software.

 

[lolfail9001]

I don't really see this Ryzen as a workstation CPU at all actually. I mean , the limited RAM support alone... what the hell am I going to do with 64Gigs?

Do stuff that is too heavy to do on laptop but urgent enough to monitor to not offload it to server with proper connectivity, obviously.

I like how everything works and I think a Naples workstation might be priced right when they get here

Honest question: would you run 2699v4 as workstation? Because that's what Naples is. Besides, i did mean workstation in wider sense than you thought of it.

 

[NKD]

Okay if you are truly trying to test this shit for 1080p gamers. Why the hell not put in a rx 480 or gtx 1060. I mean we are comparing it to real situation right. How many of the 1080p gamers throw in a fuckin gtx 1080?

I want H to do a realistic review that would be an average game scenario. cards such as rx 480 and gtx 1060 for 1080p gaming with Ryzen and then 1440p gaming and above with gtx 1080.

I know both of those situations fit me personally. As I have 1 gaming rig but I wanna update the second one in a few months just for 1080p gaming. As mine would be above.

I would want that realistic scenario that will fit that average gamer. I don't see too many people plugging in a gtx 1080 for 1080p gaming. Sure there are some. But most people are probably grabbing a 1060 or rx 480 or 470.

 

[Simplyfun]

Do stuff that is too heavy to do on laptop but urgent enough to monitor to not offload it to server with proper connectivity, obviously.

Honest question: would you run 2699v4 as workstation? Because that's what Naples is. Besides, i did mean workstation in wider sense than you thought of it.

I'm running 2x E52660 at the moment, and Naples will come in assorted sizes and speeds as well, not just one giant gob of 32 cores and nothing else.

But yes, for for what I'm doing a 2699v4 would be nice, but it's a 2700ish dollar acquisition in Canada plus I'd need a new board. I picked up the Xeons fairly cheap as new pulls from someone upgrading a server. Like I said, they'll do for now.

 

[Trimlock]

Okay if you are truly trying to test this shit for 1080p gamers. Why the hell not put in a rx 480 or gtx 1060. I mean we are comparing it to real situation right. How many of the 1080p gamers throw in a fuckin gtx 1080?

I stopped responding to this thread for a while because its pretty much the same argument over and over lol but I thought I'd respond to this.

You'd be surprised how many game on high end GPU's and only use 1920x1080 resolutions. For a combination of factors; ultra FPS gaming, max content enabling, lack of product understanding and future proofing.

 

[NKD]

I stopped responding to this thread for a while because its pretty much the same argument over and over lol but I thought I'd respond to this.

You'd be surprised how many game on high end GPU's and only use 1920x1080 resolutions. For a combination of factors; ultra FPS gaming, max content enabling, lack of product understanding and future proofing.

No actually I used to work retail. Its the other way around. all the people that wanna game for cheap are cheap. no one wants to spend 600 on a video card. I have had shit the other way around. People buying shitty cards like 100 dollar ones willing to turn down settings. I think buyers are more informed these days. If you are buying a high end system you are not likely getting 1080p screen.

Very true! I agree there are people that want every frame like their life depends on it. But more realistic situation for 1080p gaming will be rx 480 and gtx 1060.

Think about it? Do you see many reviews recommending gtx 1080 for 1080p gaming, even here at H? Its just not a realistic scenario.

 

[Trimlock]

Think about it? Do you see many reviews recommending gtx 1080 for 1080p gaming, even here at H? Its just not a realistic scenario.

I've played around with a GTX 1080 and 1070 on 1080p before and I wasn't able to generate a decent difference between the two, but on a RX 480 I could tell a difference right off the bat so there is merit. Weather its worth the price is up to the consumer, I personally don't have unique needs to game at 1080p but if I were buying right now and playing at 1080p I'd buy a GTX 1070 or a 1060. (after recent "planned" price cuts hit)

 

[NKD]

I've played around with a GTX 1080 and 1070 on 1080p before and I wasn't able to generate a decent difference between the two, but on a RX 480 I could tell a difference right off the bat so there is merit. Weather its worth the price is up to the consumer, I personally don't have unique needs to game at 1080p but if I were buying right now and playing at 1080p I'd buy a GTX 1070 or a 1060. (after recent "planned" price cuts hit)

Yea 1070 is damn good deal at 349. But it seems price drop is virtual it seems. I haven't seen any at $349. I don't think retailers are going to hit 350 if they are selling st 379.99.

I am surprised nvidia is trolling even on their own website with prices at $399 for gtx 1070 and 549.99 for 1080.

 

[Zarathustra[H]]

No I asked because I havent seen its the case.

http://techreport.com/review/31366/amd-ryzen-7-1800x-ryzen-7-1700x-and-ryzen-7-1700-cpus-reviewed/5

View attachment 18541
View attachment 18542

Why 8.3ms? That's 120fps. Way higher than anyone needs or can pass a blinded test on.

Instead of looking at average framerates, it would be more relevant to look at minimum framerates. Average framerates don't mean shit, because usually in the most intense scenes when you need your framerate the most, is when it drops the lowest.

For my uses I'd want to see minimum framerates stay above 60fps, as I game on 40 60hz. Some like/need higher framerates. 90fps is a good target here, I guess, but it is not as critical if you momentarily drop below 90.

120fps is useless meaningless overkill in 100% of all scenarios. I know lots of people think they can tell the difference. They can't. It's all placebo over 90. It's even mostly placebo over 60.

 

[lolfail9001]

Why 8.3ms? That's 120fps. Way higher than anyone needs or can pass a blinded test on.

Joke's on you, i have seen too many people claiming they can pass a "blind" test on what is 60 and what is 120 fps.
Frankly, i have never for live of me has seen 144hz monitor and even i am confident i can pass it.

 

[Zarathustra[H]]

Joke's on you, i have seen too many people claiming they can pass a "blind" test on what is 60 and what is 120 fps.
Frankly, i have never for live of me has seen 144hz monitor and even i am confident i can pass it.

I've seen a lot of people confident they can pass it too. I've never actually heard of anyone doing a true blinded test though.

People assume they can pass it based on their confirmation bias when playing around with settings, but that doesn't mean they actually can.

60 vs anything much lower than 60 is relatively straightforward.

60 vs 90 starts getting harder to detect. It is excruciating subtle.

90 vs anything above 90? My BS detector starts beeping at full capacity when anyone starts making claims that they can easily detect this difference.

 

[cybereality]

There are for sure gains above 60fps. I ran a 120Hz monitor for a while, and am on 144Hz now, and there is a noticeable difference.

Once you start getting higher, though, there are diminishing returns. Going from 60 to 90 is significant. 90 to 120 maybe not much and 120 to 144 I can't really see any difference.

 

[FrozenSun]

Great review. I still have a 2600k that runs at 4.5 ghz day in day out. Good to see that thrown in for comparison.
My question is this. I didn't see which VR setup you were testing on. The Rift or the Vive?
I'm interested in VR, and if I can get away with using my 2600k set up it will be much easier to afford the VR.
If I need a new CPU, motherboard and ram, no longer affordable.

I can't speak from personal experience, but a 2600k at 4.5GHz should be fine from everything I've read.

 

[ShuttleLuv]

^^^ In this day and age I'd say min for 2600k is 4.7 actually, 4.8 for 2500k. Just my opinion.

 

[dukesdad]

I can't speak from personal experience, but a 2600k at 4.5GHz should be fine from everything I've read.

The Rift compatibility test says no, but the Vive says its okay. Rift wants an I5-4590 or better recommended or minimum I3 6100.
My 2600 k is getting older, but I can't seem to part with it. Thought of upgrading, but see no point. Definitely not going to Ryzen.
I was an AMD fan, had a nice phenom 2 955 and was looking forward to the bulldozer with all the hype back then. It was a farce, bought the 2600k and never looked back. Now the Ryzen looks okay, but nowhere near the hype.
I've always been a fan of the underdog, but geez, not yet.

 

[Nenu]

The Rift compatibility test says no

At stock clocks.

 

[toddw]

Maxed 1080p gaming is the hardest thing I do with my rig. Everything else is MS Office or e-mail or something non-time sensitive.
As much as I hate to admit it, combined with a 1070GTX my current rig does what I want it to do. If I could change anything it would be that I didn't spend the money and get the 2600K. Remember at that time though, there was not as much gain from hyperthreading. Also it turned out the 2600K was a much better overclocker...

At this point for me... I'm thinking Ryzen 2... Until then, I just can't justify the need.... maybe when there's a ONE card solution that can push 4K maxed?
I really wanted to upgrade this cycle and this was decision time. I don't want to do SLI/Crossfire anymore.

Heck who knows my rig might even have one more GPU upgrade left in the tank. What's after 1080Ti?

 

[Presbytier]

I've seen a lot of people confident they can pass it too. I've never actually heard of anyone doing a true blinded test though.

People assume they can pass it based on their confirmation bias when playing around with settings, but that doesn't mean they actually can.

60 vs anything much lower than 60 is relatively straightforward.

60 vs 90 starts getting harder to detect. It is excruciating subtle.

90 vs anything above 90? My BS detector starts beeping at full capacity when anyone starts making claims that they can easily detect this difference.

From personal experience yes 0-60 pretty detectable 60-90 when you first start using it makes things even smoother it can be surprising to look at. 90-120 things get interesting. I personally cant tell between 90 and 120 but if the frames jump a lot between 90 and 120 I can feel that. At that high of a refresh it is more about feeling then seeing as motion sickness can set in if it jumps a lot higher than 144 I can't even see if it jumps at all though.

 

[Zarathustra[H]]

From personal experience yes 0-60 pretty detectable 60-90 when you first start using it makes things even smoother it can be surprising to look at. 90-120 things get interesting. I personally cant tell between 90 and 120 but if the frames jump a lot between 90 and 120 I can feel that. At that high of a refresh it is more about feeling then seeing as motion sickness can set in if it jumps a lot higher than 144 I can't even see if it jumps at all though.

Yeah,

If I had a choice and my hardware were fast enough to play at those framerates at my resolutions, and I had a monitor that supported those high refresh rates, I'd probably research engine fps limiters and set them to 90fps either way.

The difference above that is too small to be meaningful, and I'd strongly prefer a quieter running computer.

 

[DooKey]

Yea 1070 is damn good deal at 349. But it seems price drop is virtual it seems. I haven't seen any at $349. I don't think retailers are going to hit 350 if they are selling st 379.99.

I am surprised nvidia is trolling even on their own website with prices at $399 for gtx 1070 and 549.99 for 1080.

I bought a Zotac 1070 for 349 on Friday at Newegg. Came with a new game also.

 

[FrozenSun]

The Rift compatibility test says no, but the Vive says its okay. Rift wants an I5-4590 or better recommended or minimum I3 6100.
My 2600 k is getting older, but I can't seem to part with it. Thought of upgrading, but see no point. Definitely not going to Ryzen.
I was an AMD fan, had a nice phenom 2 955 and was looking forward to the bulldozer with all the hype back then. It was a farce, bought the 2600k and never looked back. Now the Ryzen looks okay, but nowhere near the hype.
I've always been a fan of the underdog, but geez, not yet.

As far as I can tell, the Rift test just checks your processor ID and sees if it's a model that's on their list of approved processors. I don't think it actually benches anything. A 4.5GHz 2600k is faster than a lot of the processors it will accept. I mean, looking at the IPC comparison that Kyle did a bit back only showed the 7700k being like 20% better clock for clock than the 2600k, and the 7700k is top of the line for VR. In fact, in the Ryzen review the 2600k @ 4.5GHz seemed to be fast enough for every game tested in VR.

I just ordered a Rift and am rocking an ancient 2500k, I'll let you know how it goes. I don't have super high hopes, but I'm upgrading this comp as soon as I can figure out which CPU I actually want to get.

 

[razor1]

http://www.pcgameshardware.de/Ryzen-7-1800X-CPU-265804/Tests/Test-Review-1222033/

https://translate.google.com/translate?sl=auto&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=http://www.pcgameshardware.de/Ryzen-7-1800X-CPU-265804/Tests/Test-Review-1222033/&edit-text=

Interesting enough they tested to see what will 4 cores do on the 8 core chip by disabling it in bios.

There isn't that much difference, a few % points that is it. Cutting down the core counts though have a bigger impact.

Of course this isn't ideal for testing for the CCX issues, but interesting that its still there.

 

[Presbytier]

http://www.pcgameshardware.de/Ryzen-7-1800X-CPU-265804/Tests/Test-Review-1222033/

https://translate.google.com/translate?sl=auto&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=http://www.pcgameshardware.de/Ryzen-7-1800X-CPU-265804/Tests/Test-Review-1222033/&edit-text=

Interesting enough they tested to see what will 4 cores do on the 8 core chip by disabling it in bios.

There isn't that much difference, a few % points that is it. Cutting down the core counts though have a bigger impact.

Of course this isn't ideal for testing for the CCX issues, but interesting that its still there.

Assuming the performance scales depending on how the R5 and R3 are priced that would be a compelling buy for people working within tighter budgets.

 

[razor1]

Well these motherboards are really rough, friend bought a ryan 1700x and motherboard , ended up getting a error 55 and wouldn't post, the manual says 55 is a memory error, ended up returning it, as the shop he got it from didn't have any other motherboards in stock.

Just asked him to make sure,yeah it was an asus crosshairs hero.

 

[NKD]

LOL shit we all get it, its not as fast as intel in games. 100 vs 110, or 125 vs 138. But that doesn't mean you can't game on it jeez. People have choices buy what you feel best with it. Ryzen has its value, now if I was doing other shit and gaming and I know I wouldn't care about every last frame as long as I get a fluid experience. Given that we know AMD didn't ship the final microcode until 3 weeks before launch and all the review motherboards already had the older code. It goes to show how rushed this product was. I am sure there is a lot of enhancements still to be one.

 

[NKD]

Well these motherboards are really rough, friend bought a ryan 1700x and motherboard , ended up getting a error 55 and wouldn't post, the manual says 55 is a memory error, ended up returning it, as the shop he got it from didn't have any other motherboards in stock.

Just asked him to make sure,yeah it was an asus crosshairs hero.

Basically.


just watch that video. Gamer Nexus just says it straight out the boards are a mess right now. AMD didn't ship final code until 3 weeks and they had already gotten their review samples. So I think its an unfinished product on the microcode side. Given how memory is locked down there will be lots of bios updates in the near future.

 

[razor1]

yeah probably, they launched way too soon. should waited a month or so and these types of issues would probably been ironed out.

 

[razor1]

LOL shit we all get it, its not as fast as intel in games. 100 vs 110, or 125 vs 138. But that doesn't mean you can't game on it jeez. People have choices buy what you feel best with it. Ryzen has its value, now if I was doing other shit and gaming and I know I wouldn't care about every last frame as long as I get a fluid experience. Given that we know AMD didn't ship the final microcode until 3 weeks before launch and all the review motherboards already had the older code. It goes to show how rushed this product was. I am sure there is a lot of enhancements still to be one.

Well now these are intial reviews man, not good,

just found this video for my friend he isn't the only one lol.



oddly enough this guy says 3 others were returning their ryzen products lol

So now another botched launch in 6 months first with Polaris and now with Ryzen

 

[NKD]

Well now these are intial reviews man, not good,

just found this video for my friend he isn't the only one lol.



oddly enough this guy says 3 others were returning their ryzen products lol

So now another botched launch in 6 months first with Polaris and now with Ryzen

Atleast the product itself can compete mostly. It should be fine. I remember having issues when I first jumped on intel core i hype train. It was solid product but boy I had fun with them boards.

I expected it to be 10 to 20% slower clock for clock in everything not just in games at lower resolutions.

Although i am not seeing too many posts about boards not posting here though. I think some reviewers found cores not loading as well during games.

 

[razor1]

Atleast the product itself can compete mostly. It should be fine. I remember having issues when I first jumped on intel core i hype train. It was solid product but boy I had fun with them boards.

I had issues with peripherals but nothing really with the getting the system up and going.

But any case initial buyers are beta testers for the most part. That isn't very much fun.

 

[NKD]

I had issues with peripherals but nothing really with the getting the system up and going.

But any case initial buyers are beta testers for the most part. That isn't very much fun.

I had a lot of errors with evga boards. Once I fucking got so pissed I ran a screw driver through the pins. lol. I did have some hyper threading issues initially with chips itself and sofware, so I turned it offf. That led me to mostly purchase i5 CPUs for a while because i had that mindset. But I eventually came around to it now days.

 

[razor1]

I had one eVGA board it was their first SLI board, it went bad never touched them again, and probably will never. I have never had a hardware failure when running at stock outside of eVGA.

yeah the hyperthreading issues were there for me too.

 

[NKD]

I had one eVGA board it was their first SLI board, it went bad never touched them again, and probably will never. I have never had a hardware failure when running at stock outside of eVGA.

yeah the hyperthreading issues were there for me too.

LOL I have evga micro atx now, has been rock stable. How ironic, I wouldn't have touched them ever again. But it was the cheapest x99 micro board I could get. I got it for like 150 I believe on sale. I more hated the nvidia chipset to be honest, that was biggest piece of crap I ever had. My hate was more towards the chipset it was but those were early evga days as well when it came to boards.

I gave them another shot after like 5-6 years I think.

 

[bobzdar]

http://www.pcgameshardware.de/Ryzen-7-1800X-CPU-265804/Tests/Test-Review-1222033/

https://translate.google.com/translate?sl=auto&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=http://www.pcgameshardware.de/Ryzen-7-1800X-CPU-265804/Tests/Test-Review-1222033/&edit-text=

Interesting enough they tested to see what will 4 cores do on the 8 core chip by disabling it in bios.

There isn't that much difference, a few % points that is it. Cutting down the core counts though have a bigger impact.

Of course this isn't ideal for testing for the CCX issues, but interesting that its still there.

Look at 2+2 vs 4+0, that shows you the ccx swap issue right there. 135fps vs 122fps just by using different cores. Scaling isn't linear, but I'd expect 8-10fps on 8 core with threads scheduled correctly to avoid ccx swap.

 

[cybereality]

But any case initial buyers are beta testers for the most part. That isn't very much fun.

Half the fun of being on the bleeding edge is that excitement when things actually work.

 

[razor1]

LOL I have evga micro atx now, has been rock stable. How ironic, I wouldn't have touched them ever again. But it was the cheapest x99 micro board I could get. I got it for like 150 I believe on sale. I more hated the nvidia chipset to be honest, that was biggest piece of crap I ever had. My hate was more towards the chipset it was but those were early evga days as well when it came to boards.

I gave them another shot after like 5-6 years I think.

good to know, maybe I'll try out evga again!

 

[razor1]

Half the fun of being on the bleeding edge is that excitement when things actually work.

Yeah I got too old for that lol

 

[Rod Knock]

My first theory was that gaming performance suffered because of the narrow FPU pipelines when compared to Intel. After doing some more reading and looking at other benchmarks, that's not it.

I honestly think that the cores are starved for more memory bandwidth. I doubt that higher speed memory will fix that. I think that Ryzen would have benefited from a quad channel memory controller, but I doubt that's in the cards for this CPU any time soon.

 

[FrgMstr]

The "other" project this weekend. Getting three systems fully qualified, stress tested, and ready to ship out to Brent for real world gameplay testing. No one does it as good as he does, especially not me! So I will be turning that over the master and he will be working his magic on it. 2600K at 4.5/2133 - 7700K at 5/3600 - 1700X at 4/2933