VOGONS

Common searches


AMD drops the mic

Topic actions

Reply 120 of 279, by Scali

User metadata
Rank l33t
Rank
l33t
SPBHM wrote:

they ignored clockspeed you say, yet they are able to release a CPU with 3.6GHz base and boost up to 4GHz within a 95W TDP, while Intel's 8c alternative is clocked lower (and costs 2x)

Why are you comparing to Intel?
I was specifically talking about their '51% IPC' remark, which was a comparison with Bulldozer/Piledriver. And Bulldozer runs at way more than 3.6-4 GHz. They go up to a base clock of 4.7 GHz and a turbo of 5 GHz. It was designed for higher clockspeeds at the cost of IPC, so the comparison is very skewed.

SPBHM wrote:

yes C2D IPC gain was far greater, but, I think the improvements are overall harder to achieve nowadays (look at Intel incremental approach), and we are talking about AMD

Yes, we are talking about AMD.
AMD actually took a step *back* in terms of IPC with the Bulldozer. So for them it was VERY easy to get IPC gains. In fact, they could just have re-released the Barcelona, and boast about IPC gains.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 121 of 279, by SPBHM

User metadata
Rank Oldbie
Rank
Oldbie
Scali wrote:
Why are you comparing to Intel? I was specifically talking about their '51% IPC' remark, which was a comparison with Bulldozer/P […]
Show full quote
SPBHM wrote:

they ignored clockspeed you say, yet they are able to release a CPU with 3.6GHz base and boost up to 4GHz within a 95W TDP, while Intel's 8c alternative is clocked lower (and costs 2x)

Why are you comparing to Intel?
I was specifically talking about their '51% IPC' remark, which was a comparison with Bulldozer/Piledriver. And Bulldozer runs at way more than 3.6-4 GHz. They go up to a base clock of 4.7 GHz and a turbo of 5 GHz. It was designed for higher clockspeeds at the cost of IPC, so the comparison is very skewed.

SPBHM wrote:

yes C2D IPC gain was far greater, but, I think the improvements are overall harder to achieve nowadays (look at Intel incremental approach), and we are talking about AMD

Yes, we are talking about AMD.
AMD actually took a step *back* in terms of IPC with the Bulldozer. So for them it was VERY easy to get IPC gains. In fact, they could just have re-released the Barcelona, and boast about IPC gains.

the original FX was also released at 3.6GHz max, the 4.7GHz one is a CPU with a 220W TDP released 3 years later (when the process/architecture had matured quite a bit), do you really think that's a relevant part? in any case the performance is low even compared with the 65W Ryzen (1700).

the fastest FX with the same TDP as the R7 1800x have a 3.3GHz clock (FX 8370E) and half the amount of full cores and threads,

and AMD typically was quick to start improving what they had, I think they are going to improve how they handle the 14nm process and improve the Ryzen architecture fairly well over the next few years.

not so easy, they released several revisions of Bulldozer with some IPC gains, but nothing like 50%, it took an entirely new design and 4 years,

the high clocked FXs from 2012+ with 125W TDP were all around 4GHz(turbo around 4.2), Ryzen starting it's life at 3.6GHz(turbo around 4) is positive I think, the sacrifice in clock doesn't look to significant when you add the IPC gain.

also, the 52% gain is over newer revisions of Bulldozer, not the 2011 one
http://i.imgur.com/zavUwMV.jpg

Reply 123 of 279, by Scali

User metadata
Rank l33t
Rank
l33t

Well, reviews are out now:
http://www.anandtech.com/show/11170/the-amd-z … -1700x-and-1700
https://arstechnica.com/gadgets/2017/03/amd-ryzen-review/
http://www.tomshardware.com/reviews/amd-ryzen … x-cpu,4951.html
http://hexus.net/tech/reviews/cpu/102964-amd- … 1800x-14nm-zen/
https://www.extremetech.com/gaming/245204-amd … g-achilles-heel

As expected, not quite up to Intel's IPC and single-threaded performance, therefore not that interesting for gaming and such.
They are mainly interesting for 2 reasons:
1) Intel's 8-core/16-thread offerings are currently priced much higher (but will no doubt drop considerably after today's launch).
2) Intel's 8-core/16-thread CPUs are still based on the outdated Broadwell architecture.

So AMD did an excellent job of going for Intel's weak spot.
AMD's own weak spot is also obvious in some of the benchmarks: sometimes it gets beaten by much cheaper/older/less powerhungry Intel CPUs with just 4 cores.
So I don't think the 4-core variations of Zen will be a big success. The 8-core versions at least have good multithreading performance to offer.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 124 of 279, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++

Kyle Bennett noted in his review that the gaming performance issue seems to be largely confined to low resolution test sets and that this performance delta fades somewhat at higher resolution or graphical detail settings. I'm not sure whether this is just the trade between GPU and CPU power at work (which is good, as it means that the peripheral logic is efficient) or perhaps an optimization issue. Anyway, the conclusion was basically: good at gaming, great at anything multithreaded.

Ryzen IPC appears to be quite strong, but it clocks somewhat less than Intel's highest clocked parts. What I'm really impressed to see is the power draw benchmarks.

All hail the Great Capacitor Brand Finder

Reply 125 of 279, by Scali

User metadata
Rank l33t
Rank
l33t
gdjacobs wrote:

Kyle Bennett noted in his review that the gaming performance issue seems to be largely confined to low resolution test sets and that this performance delta fades somewhat at higher resolution or graphical detail settings. I'm not sure whether this is just the trade between GPU and CPU power at work (which is good, as it means that the peripheral logic is efficient) or perhaps an optimization issue.

It's pretty simple: Higher resolution and/or detail settings make the game more GPU-limited, frame rates drop, and therefore the CPU becomes less of a factor in performance.
So it's not like the CPU gets better in those circumstances.

gdjacobs wrote:

Anyway, the conclusion was basically: good at gaming, great at anything multithreaded.

Well I suppose I read different reviews then.
Especially the Ars Technica review is pretty blunt about it: Ryzen 7 is not for gaming.
A 4-core Intel is faster AND much cheaper in games. Even the ones that are 'great multithreading' games.

I read an even more painful detail: AMD wants to talk about how streaming your game live is better with more cores.
Theoretically that's true. The reviewer measured that a Core i7 7700k dropped about 18 fps, while the Ryzen only dropped 3 fps.
So far so good. The flipside however: The 7700k had more than a 15 fps margin to begin with, so even when streaming game video it was still faster, despite taking a larger performance hit.

Moar coars is so 2011.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 126 of 279, by swaaye

User metadata
Rank l33t++
Rank
l33t++

I like Tech Report's frame time results. You get an idea of overall fluidity. Ryzen looks better than Bulldozer but yeah Intel is still the way to go.
http://techreport.com/review/31366/amd-ryzen- … cpus-reviewed/5

Seems like it's an interesting chip if you can really use all those threads.

Reply 127 of 279, by Scali

User metadata
Rank l33t
Rank
l33t
swaaye wrote:

Seems like it's an interesting chip if you can really use all those threads.

Yea, it's the classic AMD recipe: an 8-core CPU with almost the performance of an Intel, at a fraction of the price.
But pricing with 2-core and 4-core CPUs is much tighter, and those CPUs use a more modern architecture than the 8-core CPUs. So I don't really see how AMD is going to make Ryzen an obvious choice in that market.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 128 of 279, by DosFreak

User metadata
Rank l33t++
Rank
l33t++

Hopefully they keep up momentum so when I'm looking to for a new CPU in a minimum of 5 years they'll be an option. I don't purchase any new technology until a year after release anyway and this is AMD so the results aren't suprising.

So far seeing commenters complain about variance in benchmarks in reviews due to SMT enable/disabled, possibly issues with motherboards and memory speed limitations and blaming game benchmarks for not being optimized for AMD processors. heh.

I've also been more and more concerned with minimum over average FPS as the years have gone and so far I'm seeing one review saying the scores are much lower. Haven't had a chance to read all the reviews yet.

Don't give a shit about streaming or video conversion since I makemkv all my movies so that does nothing for me and my desktop is purely a gaming machine.

How To Ask Questions The Smart Way
Make your games work offline

Reply 129 of 279, by Standard Def Steve

User metadata
Rank Oldbie
Rank
Oldbie

Well, the gaming performance is kinda disappointing. I think my 4.5GHz 4930K (Ivy-E with quad DDR3-2400) might actually beat this thing in games. The single-threaded performance drop with SMT enabled is also a little concerning.

It's a beast at video encoding, right up there with the 8c/16t 6900K. As I do a fair amount of x264 and 265 encoding these days, I'm still quite tempted to buy the cheaper 1700 and overclock it to around 4GHz. The two additional cores and AVX2 should easily outperform my 6c/12T i7.

94 MHz NEC VR4300 | SGI Reality CoPro | 8MB RDRAM | Each game gets its own SSD - nooice!

Reply 130 of 279, by Jade Falcon

User metadata
Rank BANNED
Rank
BANNED

If what I'm seeing is true, a server version of this chip will be killer. Intel will have its hands full with the server side of things if amd could get there foot in the door. Unfortunately amd has never been able to mass market server cpus like intel. Or at least not in my experience.

I like amd for there server cpus, but hardly any enterprise business buys them like intel cpus.

Reply 131 of 279, by snorg

User metadata
Rank Oldbie
Rank
Oldbie

My primary interest in this CPU is nothing to do with games, I want it for CG/effects/animation etc. Assuming the 16 and 32 core parts are not ridiculously expensive (I have hope given the 8 core parts start at roughly $300) then I look forward to being able to put together a 32 to 64 core box on the cheap, basically a renderfarm in a box. Especially if they have 65w server parts, I could convert my Xeon system to winter-only usage and make the Ryzen box my main workstation. So yeah, I'm very interested in this chip. I think Google and Facebook will be, too. Also Amazon. If you're looking for cheap Xeons get ready to see a lot of them in the next 2 years as various large datacenters migrate, assuming the performance is as good as I think it will be.

Reply 132 of 279, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++
Scali wrote:

It's pretty simple: Higher resolution and/or detail settings make the game more GPU-limited, frame rates drop, and therefore the CPU becomes less of a factor in performance.
So it's not like the CPU gets better in those circumstances.

Agreed, but the platform becomes more relevant.

Anyway, there may be something more. We'll have to see what hashes out over the next week or two.
https://www.pcper.com/reviews/Processors/AMD- … ing-Performance

Scali wrote:
Well I suppose I read different reviews then. Especially the Ars Technica review is pretty blunt about it: Ryzen 7 is not for ga […]
Show full quote

Well I suppose I read different reviews then.
Especially the Ars Technica review is pretty blunt about it: Ryzen 7 is not for gaming.
A 4-core Intel is faster AND much cheaper in games. Even the ones that are 'great multithreading' games.

I read an even more painful detail: AMD wants to talk about how streaming your game live is better with more cores.
Theoretically that's true. The reviewer measured that a Core i7 7700k dropped about 18 fps, while the Ryzen only dropped 3 fps.
So far so good. The flipside however: The 7700k had more than a 15 fps margin to begin with, so even when streaming game video it was still faster, despite taking a larger performance hit.

Moar coars is so 2011.

He was pretty clear that Kaby Lake is a better processor for gaming because of higher single thread performance, but the performance gap (~15%) is never large enough that Ryzen can't game well, especially as higher resolution reduces the performance delta. Content creation and video editing is a win, so if that factors into your purchasing decision, Ryzen might be the processor you want.

All hail the Great Capacitor Brand Finder

Reply 134 of 279, by Kreshna Aryaguna Nurzaman

User metadata
Rank l33t
Rank
l33t

Straight from Tech Report. It seems Ryzen has the best price to performance ratio for non-gaming purpose, which is good enough for me. To me, GPU is always more important than CPU for gaming, since I always aim for the highest Anti-Aliasing performance possible. And AA is where GPU matters more than CPU.

value-productivity.png

Never thought this thread would be that long, but now, for something different.....
Kreshna Aryaguna Nurzaman.

Reply 135 of 279, by PhilsComputerLab

User metadata
Rank l33t++
Rank
l33t++

I don't game much these days, mostly on retro computers. The performance for video work looks great, but I won't buy first release gear 😀 Waiting for the refresh and all kinks to reveal themselves / get addressed.

Regarding gaming one review noted that the AMD was smoother, though lower FPS. The Intel would have little skips or stutters. So that's something interesting to look into.

YouTube, Facebook, Website

Reply 136 of 279, by Scali

User metadata
Rank l33t
Rank
l33t
gdjacobs wrote:

He was pretty clear that Kaby Lake is a better processor for gaming because of higher single thread performance, but the performance gap (~15%) is never large enough that Ryzen can't game well, especially as higher resolution reduces the performance delta. Content creation and video editing is a win, so if that factors into your purchasing decision, Ryzen might be the processor you want.

Problem is, those Intel CPUs that are better at gaming are much cheaper than the Ryzen models introduced so far.
I say it's useless to compare those CPUs. We'll have to wait for the 6-core and 4-core Ryzen models to get a better idea of where the Ryzen stands when the prices are comparable.

Speaking of prices, Intel's prices aren't set in stone, so I don't expect AMD to retain the current price-advantage in the 8-core market for very long, and Intel may even reduce their prices for 6-core and 4-core CPUs to make life for Ryzen more difficult.
So we'll have to re-evaluate Ryzen in a few weeks or so, when we know how Intel has responded price-wise.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 137 of 279, by Scali

User metadata
Rank l33t
Rank
l33t
PhilsComputerLab wrote:

I don't game much these days, mostly on retro computers. The performance for video work looks great, but I won't buy first release gear 😀 Waiting for the refresh and all kinks to reveal themselves / get addressed.

Yes, various reviewers mentioned that they had problems getting the DDR4 modules working at their advertised speeds, and had to run them at lower speed to get the system stable.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 138 of 279, by Kreshna Aryaguna Nurzaman

User metadata
Rank l33t
Rank
l33t
PhilsComputerLab wrote:

I don't game much these days, mostly on retro computers.

Same here. My dedicated gaming rigs are build around DOS and Windows 9x.

For (relatively) modern games, I don't have dedicated gaming rig. My main rig is 32-bit Windows 7, that works simultaneously as gaming rig, working rig, and audiophile PC. I mostly work with Office (mostly PowerPoint) while having probably a dozen of browser tabs, while listening to music on the said PC. As such, multi-threaded performance is important to me.

Single-threaded performance, on the other hand, is not really important, since the games I play are mostly the oldest Direct3D/OpenGL games possible that can still run on the said rig. Examples are Crimson Skies and Neverwinter Nights. I just like to try such older games on modern PC, enabling the highest AA possible to see how would it look on such games. As such, GPU is far more important to me than CPU. That's particularly true for Crimson Skies, which runs too fast on a 2.4 GHz i5 where there aren't much textures around (like night mission). As such, my building philosophy for main rigs are something like this: "fastest GPU possible, combined with CPU whose single-threaded performance is slow enough for old games, but multi-threaded performance is good enough for office works."

Yeah, I know my case is extreme, but I know a lot of people who prioritize GPU over CPU.

Never thought this thread would be that long, but now, for something different.....
Kreshna Aryaguna Nurzaman.

Reply 139 of 279, by Scali

User metadata
Rank l33t
Rank
l33t
Kreshna Aryaguna Nurzaman wrote:

For (relatively) modern games, I don't have dedicated gaming rig. My main rig is 32-bit Windows 7, that works simultaneously as gaming rig, working rig, and audiophile PC. I mostly work with Office (mostly PowerPoint) while having probably a dozen of browser tabs, while listening to music on the said PC. As such, multi-threaded performance is important to me.

Not really.
It's not like you have more (active) threads just because you have multiple programs/tabs open. In fact, most browsers will put invisible tabs on idle.
That's just the multicore-myth right there.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/