AMD Ryzen (Summit Ridge) Benchmarks Thread (use new thread)

[Insert_Nickname]

I remember back in the day that i had to "fight" "people" who argued that E-450 was better in games than an I3 with HD3000...

Why bother? Both were equally incapable of anything more then absolute lowest settings on lowest resolution. If you're OK with that, there is essentially no difference. Brazos did have better drivers though.

Anyway, Brazos was a piece of crap and it sold a lot because OEM used it EVERYWHERE, from netbooks to 17" notebooks. It was better than Atoms but that was it, and OEMs placed where they never meant to be.

To be fair to the various Brazos derivatives they were decent when launched. In 2011. Using x86 software. They did not hold up well past that, and have aged extremely badly, particularly with x64 software. And OEMs loved to burden them with x64 versions of Windows.

The low-clocked C-series really were abominations, that should never have left the factory. That is not something I'm stating lightly. The damage they have done to the perception of the PC platform for ordinary users is incalculable.

 

[Shivansps]

Why bother? Both were equally incapable of anything more then absolute lowest settings on lowest resolution. If you're OK with that, there is essentially no difference. Brazos did have better drivers though.



To be fair to the various Brazos derivatives they were decent when launched. In 2011. Using x86 software. They did not hold up well past that, and have aged extremely badly, particularly with x64 software. And OEMs loved to burden them with x64 versions of Windows.

The low-clocked C-series really were abominations, that should never have left the factory. That is not something I'm stating lightly. The damage they have done to the perception of the PC platform for ordinary users is incalculable.

The problem, back in the day was, I was a owner of a MSI U230 (Neo X2 L335 and 780G mobile), them i changed to HP DM1z (E-350) belive me, the CPU was SLOWER, not much slower, but it was, and in games, the E-350 was 100% CPU limited, killing most of the jump in IGP perf, especially considering i was able to overclock the 780G on the MSI U230.
Any I3, even a U, and probably even Pentiums where a lot better in CPU, thus better in games, an I3 M whould run cicles around it, for example E-350 was unable to run GW2 at any detail or resolutions, I3-2310M could run it at playable fps. But trying to tell people that IGP alone was not enoght was HARD, i whould not be suprised if E-350 perf in games was worse than Pentiums with HD2000, in fact im 90% sure the Pentium whould perform better with DC ram.
There was also the problem of SC ram limiting the E-350 IGP, 780G mobile had DC.

Anyway, the problem was, Brazos was intended to target products with Atoms, Neos and maybe Celeron/Pentium U SB... but OEM started to put them in mainstream 14" and 15" notebooks, and more, in producs where AMD Phenom 2 mobile or Pentium/I3 SB M was used, that was terrible, not to mention they also started using C-Brazos, Lenovo sold C-50 14" notebooks until about 2 years ago... They should have never been in larger than 12-13" devices.

 

[AtenRa]

They should have never been in larger than 12-13" devices.

Agreed and i would even say not above 11.6".

 

[Shivansps]

I was talking about Brazos in general, E-350 was not too bad for 12" netbook and 13" ultrabooks, it was OK for that, C-Brazos should never have been in larger than 10.1" devices, 11.6" could handle E brazos just fine, and Single Core Brazos should have never existed.

 

[AtenRa]

I was talking about Brazos in general, E-350 was not too bad for 12" netbook and 13" ultrabooks, it was OK for that, C-Brazos should never have been in larger than 10.1" devices, 11.6" could handle E brazos just fine, and Single Core Brazos should have never existed.

I was talking about C models for up to 11.6"

 

[KTE]

^Yea I would agree, that was my painful assessment back then too.

In the mobile world, even when AMD had something niche and competitive, their strategy ruined it.

I have an E350 right infront me in a 15.6", and it runs dog slow with basic tasks.

I also have a 3227U in a 13" chassis right next to it (and a C50 in 10", a SU7300 in 11.6", a 3667U in 12.5", a 5300U in a 12.5" and a 2310M in 12.5").

The latter runs rings around the former. In a very old Q3 port I sometimes play (Urban Terror, and CS:S actually), it's perfectly playable with the IVB but completely unplayable with the Zacate solution. Just as an example.

To compare, the ATI HD 5450 played the games flawlessly.

The DP latency on the Zacate solution is also extremely high, and the mouse cursor/window scrolling or movement lag.

Very low clocks or very low IPC ruins a solution.

Sent from HTC 10
(Opinions are own)

 

[krumme]

The entire zakata die was 75mm2 on cheap leaking 40nm while a sb 2c was about twice as big on 32nm. The bobcat cpu was 4.6mm2 or less than half size of 45nm Atom. With a 50% ipc lead btw.
http://www.chip-architect.com/news/2010_09_04_AMDs_Bobcat_versus_Intels_Atom.html
Comparing zakate to SB is utterly idiotic from an economic perspective. I fail to see how its relevant on a technical level either.

Now zen is probably far closer to core than the new Atom line but bocat was surely more squarely on Atom size. But mm2 is interesting as this core or a derivative of it on 10 nm will probably have to fit in a new console, or fight for it at least. And it seems borderline imo at that as i can tell?

 

[Sven_eng]

Brazos was meant to compete with Atom, which it did and more. Atom sales crashed only weeks after the E-350 launched. It was OEM's who put it in craptops, not AMD.

If you consider the updated core lives on still in the PS4 and X1, it's probably one of the most successful x86 architectures ever in terms of units sold.

 

[Shivansps]

I was talking about C models for up to 11.6"

But 11.6" form factor could handle E-350 18W just fine, there was not really a need to use C-50s there, in fact, that was the start of the problem, OEMs started using C Brazos on products intended for E Brazos, and E Brazos somewhere else.

 

[AtenRa]

AMD RYZEN - Trademark Details

https://trademarks.justia.com/871/15/amd-87115584.html

 

[Shivansps]

The entire zakata die was 75mm2 on cheap leaking 40nm while a sb 2c was about twice as big on 32nm. The bobcat cpu was 4.6mm2 or less than half size of 45nm Atom. With a 50% ipc lead btw.
http://www.chip-architect.com/news/2010_09_04_AMDs_Bobcat_versus_Intels_Atom.html
Comparing zakate to SB is utterly idiotic from an economic perspective. I fail to see how its relevant on a technical level either.

Now zen is probably far closer to core than the new Atom line but bocat was surely more squarely on Atom size. But mm2 is interesting as this core or a derivative of it on 10 nm will probably have to fit in a new console, or fight for it at least. And it seems borderline imo at that as i can tell?

Its relevant because OEM started to place it in products that compited with SB, where Atoms had no products, when you place a E-350 in a 14 to 17" notebook, you can be sure you will have a hard time trying to find a Atom notebook to compare it with.

Cheap notebooks was the reason of why it sold so much, and why people argued that it was better than SB Pentiums and Core I3 for games.

 

[jpiniero]

AMD RYZEN - Trademark Details

https://trademarks.justia.com/871/15/amd-87115584.html

That's a cheesy enough name to make me think that is what SR is going to be called.

 

[AtenRa]

Cheap notebooks was the reason of why it sold so much, and why people argued that it was better than SB Pentiums and Core I3 for games.

E-450 (HD6320) was better than SB Pentiums in games using iGPUs (HD2000) , against i3 (HD3000) mobile it was game depended. Some games were faster on E-450 and some on the i3. When the game needed more CPU then the i3 was the better choice. The price was also in favor of the E-450 if i remember correctly.

 

[Thunder 57]

Are you sure? I could very well be wrong on that. Makes sense since socket 940 for the first desktop A64 was the same socket as Opteron.

Sorry, didn't mean to sound like a know it all, I know that can be annoying. Especially when it really isn't important anymore. But yes, the Opteron came out in February or March I want to say, and the A64 followed in September. I remember Anand had a series of articles predicting the A64 performance based on the Opteron platform. Had I not remembered that I would have no idea which had come first.

 

[krumme]

E-450 (HD6320) was better than SB Pentiums in games using iGPUs (HD2000) , against i3 (HD3000) mobile it was game depended. Some games were faster on E-450 and some on the i3. When the game needed more CPU then the i3 was the better choice. The price was also in favor of the E-450 if i remember correctly.

Many of the games you would play on this type of machine would be severely st limited on e350. Eg tf2. Core i3 was in another league even for gaming imo.

I bought a hp 11.6 back then with a e350 for one of the kids and considered a thinkpad 11.6 that also ran an e350. Both machines were like at least 20% more expensive with a i3 meaning the zakate was probably a quarter of the cost of core. It was completely different price segments.

E350 were dog slow but compared the the atom we had for travel before it was a F1 rocket. Lol.

I am still amazed by the power/mm2 you get in the small core lines from either camp. Have a bt nas and a bt spotify server.
Will love a quad zen nas and 10G. Wroooom.

So the possibility of zen beeing smallish is to my taste. Fill the house with it.

 

[MajinCry]

The HD 2000 isn't good for anything. In order to get morrowind running on the lowest details at 20fps, I had to resize all the game's textures to 32x32, and set the resolution to 640x480.

Did the same for New Vegas, even disabled all LOD, imagespace effects, resized the textures to 32x32, and lowered the res to 320x240. Didn't even break past 12fps.

My Thinkpad T43's ati x300 ran games better than that.

 

[VirtualLarry]

My Thinkpad T43's ati x300 ran games better than that.

I think that the x300 was what I had on an old HP laptop with XP. Played Unreal Tournament (original) really well at 640x480. I'm having some trouble remembering the CPU. It could have been a Pentium M, or maybe it was an AMD Sempron.

 

[moonbogg]

AMD RYZEN - Trademark Details

https://trademarks.justia.com/871/15/amd-87115584.html

Oh cool, I think I get it. Its like, AMD has been dead for years, but with their new CPU they will have risen. So, by saying RYZEN, it sounds like risen, but still includes the CPU's name (ZEN). Did I get that right? This thing better not be another bulldozer, because the internet could have some fun with this name. Like, AMD was braZEN enough to think they could attack Intel. Or AMD's market share has just froZEN at absolute zero. It could be fun.

 

[waffleironhead]

AMD RYZEN - Trademark Details

https://trademarks.justia.com/871/15/amd-87115584.html

Look at the other ones in progress.
GROK
JORO
VARA
THREADRIPPER

 

[majord]

Oh god. those trademarks can only mean one thing..

Another Heros style cartoon incoming

 

[bakyt115]

http://www.amd.com/en-us/innovation...m_source=301&gclid=CJu1gbWnutACFdc9gQoda2gE9A

fairy tales end

40% more performance than previous AMD core, without increasing power.*

 

[.vodka]

AMD RYZEN - Trademark Details

https://trademarks.justia.com/871/15/amd-87115584.html

Interesting name... Ryzen SR3/5/7... hmm... it has some ring to it. It could work.


Now, was there anything wrong with Phenom III?

40% more performance than previous AMD core, without increasing power.*

They don't seem to make their minds up. You have that, and a little below..

AMD’s high-performance x86 Core “Zen” architecture delivers an unprecedented 40 percent improvement in instructions-per-clock cycle over the previous generation AMD core, without increasing power.*

So, AMD defines performance as IPC, same as we simple forum mortals define IPC as perf/MHz?

If so, XV *1.4 is around SB/IVB performance, as expected. Not that there's anything wrong with that, if the price is right. We have this confirmed, and then we have that blender benchmark where 8C16T Zen equals 8C16T BW-E... this is all over the place. SB baseline performance and up to BW-E performance depending on the case? Could be, there's no "up to" anywhere to be seen.

 

[Shivansps]

E-450 (HD6320) was better than SB Pentiums in games using iGPUs (HD2000) , against i3 (HD3000) mobile it was game depended. Some games were faster on E-450 and some on the i3. When the game needed more CPU then the i3 was the better choice. The price was also in favor of the E-450 if i remember correctly.

Ill find that hard to belive since, the igps performed nearly the same, the E-350/E-450 IGP was an HD5450 with lower core clock and shared memory bandwidth, althought some games way work horribly on Intel IGPs, the HD3000 performed very very close to desktop HD5450, and CPU performance difference was just huge, on Pentiums with HD2000 that could have been the case of some games working better, some worse, but HD3000 on Core i3 was definatelly better.

Note: Always talking with Dual Channel Enabled on Intel cpus...

EDIT: you need to remember that E-350/E-450 was never meant to be compared with SB/HD3000, OEMs forced that, the Intel solution was just better, AMD LLano mobile was the one to compite with SB/HD3000.

 

[majord]

http://www.amd.com/en-us/innovation...m_source=301&gclid=CJu1gbWnutACFdc9gQoda2gE9A

fairy tales end

40% more performance than previous AMD core, without increasing power.*

yes, there's also a slide from Hotchips saying the same thing.

 

[LTC8K6]

*Based on AMD internal testing