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

[frozentundra123456]

So you are trying to say Zen runs at the base clock with all cores loaded? That seems highly unlikely.

Even if that is true, it is still the maximum performance available without overclocking. It just makes the ipc improvement better and the clockspeed deficit even worse. The bottom line is still that the 8370 is running at 3.2, and can clearly run at 4.0 on all cores at stock, giving it an additional 25% available performance compared to Zen.

 

[The Stilt]

So you are trying to say Zen runs at the base clock with all cores loaded? That seems highly unlikely.

Even if that is true, it is still the maximum performance available without overclocking. It just makes the ipc improvement better and the clockspeed deficit even worse. The bottom line is still that the 8370 is running at 3.2, and can clearly run at 4.0 on all cores at stock, giving it an additional 25% available performance compared to Zen.

Are you talking to me?

 

[Erenhardt]

So you are trying to say Zen runs at the base clock with all cores loaded? That seems highly unlikely.

Even if that is true, it is still the maximum performance available without overclocking. It just makes the ipc improvement better and the clockspeed deficit even worse. The bottom line is still that the 8370 is running at 3.2, and can clearly run at 4.0 on all cores at stock, giving it an additional 25% available performance compared to Zen.

In a form of 220Watt binned SKU. Right. There are very slim chances this ES is 220watt part clocked to its max freq.

 

[frozentundra123456]

In a form of 220Watt binned SKU. Right. There are very slim chances this ES is 220watt part clocked to its max freq.

8370 has a base clock of 4.0 ghz, as does the 8350, in a 125 watt TDP. Are you claiming they actually require 220 watts to maintain base clock?

Are you talking to me?

No.

 

[Doom2pro]

As I've said before Zen has slightly different method the generate the CPU clocks, which leads to somewhat unusual operating frequencies. The clock frequencies posted by "AMD Polaris" match this pattern. Also the CPU name string displayed by AotS is a real deal. And since the stepping identifiers and all the other information in the SKU name string is design specific, it cannot be forged either.

So your saying it's extremely hard to copy details that are publicly available (as you said from AMD Polaris) to make a hoax? How so? About the SKU, if someone KNOWS the SKUs but doesn't have the actual hardware, why can't they inject those into the benchmarks to make them appear more legit?

You say it's extremely difficult, I don't see how... Some of the information is public knowledge, not exactly state secrets here... The SKU could be had by someone who doesn't have physical access to an ES and just as easily spread as the ES leaks.

 

[The Stilt]

So your saying it's extremely hard to copy details that are publicly available (as you said from AMD Polaris) to make a hoax? How so? About the SKU, if someone KNOWS the SKUs but doesn't have the actual hardware, why can't they inject those into the benchmarks to make them appear more legit?

You say it's extremely difficult, I don't see how...

You can forge the Mona Lisa if you're skilled (knowledgeable in this case) enough, but what are the odds?

Also if the results would show Zeppelin matching a 5960X, we most likely wouldn't be having this conversation, at least in the same extent.

 

[Doom2pro]

You can forge the Mona Lisa if you're skilled (knowledgeable in this case) enough, but what are the odds?

Also if the results would show Zeppelin matching a 5960X, we most likely wouldn't be having this conversation, at least in the same extent.

So taking the information from AMD Polaris, and the SKU itself and baking it into a benchmark is equivalent to forging a painting? What stops anyone from taking this quite easily obtainable information and making a fake benchmark?

Btw, how many people do you think have lists of these SKUs for products that are being released this year and next year? Surely they aren't locked in a safe at AMD HQ, they have to be making their rounds by now...

I don't see how anyone can't just take this information and then forge a benchmark with them... I mean they practically obfuscated the fake results by using a video game benchmark that doesn't scale well with more than 8 threads instead of a half decent CPU benchmark. This wreaks of fakery... I'm not convinced, sorry.

 

[Dresdenboy]

There are several things in the information received from these leaks and all of them are extremely hard, if not impossible to forge. As I've said before Zen has slightly different method the generate the CPU clocks, which leads to somewhat unusual operating frequencies. The clock frequencies posted by "AMD Polaris" match this pattern. Also the CPU name string displayed by AotS is a real deal. And since the stepping identifiers and all the other information in the SKU name string is design specific, it cannot be forged either.

I'd say both of the leaks we've had so far are real, the only question is if the actual scores in AotS are. Most likely they're too, since the leak came from China / Taiwan which at this point makes are perfect sense.

One point I saw at one forum was, that memory speeds have a significant effect on AotS framerates. But as we didn't get them with the benchmark scores, this is just another base for speculation.

Regarding ES OPN encoding, one guy at P3DNow! posted this interesting link:
http://www.cpu-world.com/info/id/AMD-ES-identification.html

Also if the results would show Zeppelin matching a 5960X, we most likely wouldn't be having this conversation, at least in the same extent.

This would stir up the opposite direction of arguments (people love to discuss): "AotS might just be that one cherry picked benchmark for Zen due to good scaling with threads. Reality will be different."

I think this way it's even better, because it dampens expectations, should help with regaining objectivity (if lost), and might even increase the chances for positive surprises in some apps (reality > expectations).

 

[lifeblood]

Based on these leaks, if 2.8/3.2 are the final clocks then we're in for another Bulldozer scale disappointment. 40% IPC improvements quickly loose their luster if your clock speed is ~25% less.

I'm trying to take this leak with a grain of salt but past disappointments from AMD don't give me a lot of confidence.

 

[Azuma Hazuki]

I'm betting I've bought more AMD equipment than you have.

My opinions bothering you to such a degree says more about you than they say about me.

What is this schoolyard baiting redacted, Phynaz? You've had a huge chip on your shoulder since before I've joined here and you seem to take any opportunity you can to bash on AMD. This is foolish in the extreme; you'd be paying four figures for God-awful Itanic if it weren't for them, your precious Intel has been poisoning the market for over a decade and getting away with it, and yet you sit here and act like this?

I'm really tired of your immaturity and your obvious trolling, and am mystified as to why you haven't gotten more red cards than an entire soccer league on crystal meth.

Further point: testing what is obviously a server-based architecture with a game benchmark is window-licking retarded. Know what kind of bench would be a hell of a lot more relevant? Compiler tests. How fast does Zen compile the Linux kernel, on one core only and with all 8 cores/16 threads blazing? Compare to everything from Haswell i3s up to Broadwell-E. THAT gives us some idea of this thing's real-world usefulness.

 

[Soulkeeper]

I really need to build a new work computer and have been waiting for zen.
It's hard with all the intel options out there and my APU starting to feel slow with poorly threaded java apps.
Is it confirmed that we won't be able to build a zen system before xmas ?

 

[Phynaz]

What is this schoolyard baiting bullshit, Phynaz? You've had a huge chip on your shoulder since before I've joined here and you seem to take any opportunity you can to bash on AMD. This is foolish in the extreme; you'd be paying four figures for God-awful Itanic if it weren't for them, your precious Intel has been poisoning the market for over a decade and getting away with it, and yet you sit here and act like this?

I'm really tired of your immaturity and your obvious trolling, and am mystified as to why you haven't gotten more red cards than an entire soccer league on crystal meth.

Further point: testing what is obviously a server-based architecture with a game benchmark is window-licking retarded. Know what kind of bench would be a hell of a lot more relevant? Compiler tests. How fast does Zen compile the Linux kernel, on one core only and with all 8 cores/16 threads blazing? Compare to everything from Haswell i3s up to Broadwell-E. THAT gives us some idea of this thing's real-world usefulness.

Thanks for proving my point.

Now that you have called out both a member and the moderators are you ready to pull up your big boy pants and get back on topic?

What do you expect shipping clocks to be for Zen?

 

[Arachnotronic]

I really need to build a new work computer and have been waiting for zen.
It's hard with all the intel options out there and my APU starting to feel slow with poorly threaded java apps.
Is it confirmed that we won't be able to build a zen system before xmas ?

Nothing confirmed but really just buy an Intel chip at this point. Summit Ridge isn't going to be a better solution than the 6700K for your poorly threaded apps, which is now selling for just $320 on Newegg.

 

[Azuma Hazuki]

Thanks for proving my point.

Now that you have called out both a member and the moderators are you ready to pull up your big boy pants and get back on topic?

What do you expect shipping clocks to be for Zen?

#1: Big girl pants, thanks.
#2: Cry harder.
#3: Projection is an awful thing; reflect on yourself and your behavior. And I don't think someone who deliberately chose that juvenile avatar ("Lookit, guys, it's a hand and a mouse BUT IT'S TOTALLY LADYBITS HURR HURR I'M SO EDGY LIKE A RAZOR WITH 8 BLADES AND ONE OF THEM'S VERTICAL!") has any right to lecture other people on maturity...
#4: Regarding clock rates, I dunno *shrug* I'm fine with them as they are, wouldn't mind seeing higher. There aren't (usually) bad products so much as bad prices.
#5: The only "point" you have is on the top of your head (see #3).
#6: Cry harder still.

You forgot #7 which is get yourself an infraction for member callout which you have gloriously done.

Moderator Aigo.

 

[VirtualLarry]

Somebody's getting a time out soon.

Seriously, though. If Zen is 8C/16T, and Sandy Bridge IPC, that's not a bad thing(*). Really, at that point, it comes down to pricing, both unit SKU, and TCO, which involves some performance/watt calculations. (Which I believe Phynaz is involved in something related to that professionally, no?)

(*) For general usage, I mean. Gaming benefits from high-IPC / high-clocked CPUs, and much less from multi-core, at least thus far. (See: games that "require" a quad-core, run fine on a quad-threaded dual-core.)

 

[deasd]

I really need to build a new work computer and have been waiting for zen.
It's hard with all the intel options out there and my APU starting to feel slow with poorly threaded java apps.
Is it confirmed that we won't be able to build a zen system before xmas ?

I searched but found nothing about java benchmark. Reason why java run slow is not just because of poorly threaded, java is already obsoleted and there's very few optimization on it. I think there's no need to consider a new rig only for java. IF you need, i3 or A10 which close to 4Ghz clock might help. As for Zen, it might not widely available before January.
Don't spend more money on inappropriate hardware.

 

[cbn]

Deleted.

 

[Insert_Nickname]

It's hard with all the intel options out there and my APU starting to feel slow with poorly threaded java apps.

IF you need, i3 or A10 which close to 4Ghz clock might help.

If you just need single thread performance cheaply, the i3 6300* is a beast. Quite cheap too.

*or 6320, but the extra 100MHz is hardly worth the extra cost.

 

[Soulkeeper]

thanks for the replies
basically i'm torn between 4 options:

1) wait for zen
2) i7-5775C (so I can reused my fast ddr3)
3) i7-6700
4) E3-1275 v5 + tons of ecc memory (just cause)

When I built this APU in 2011 I told myself i'd wait for AMD to update their desktop platform, i'm still waiting

I guess I should be in a different thread. Sorry for the interuption.

 

[Erenhardt]

If you just need single thread performance cheaply, the i3 6300* is a beast. Quite cheap too.

*or 6320, but the extra 100MHz is hardly worth the extra cost.

Hardly... pentium anniversary Ed. at 4.5Ghz smokes both. /OT

Do we know if 4c/8t part will be a cut down die? I guess if they fab it at glofo they will probably save some money on designing 'small die' Zen and use defective or 'artificially defective' dies. Anyone in for unlocking to 8c/16t from 4c/8t a'la old athlon x2 to phenom x4 ? They surely can 'waste' a lot of GF silicon since they will have to pay for unused capacity

I hope they get the APUs right and are not afraid to pump some HBM into the mix. If they release AM4 apu with HBM at 125W or more I'm buying it day 1.

 

[blublub]

If you guy search a little you find that AMD's 8350 engineering sample was clocked at 3ghz and final was 4ghz, so there is quite some room for improvement here...

 

[Insert_Nickname]

Hardly... pentium anniversary Ed. at 4.5Ghz smokes both. /OT

OT; You don't OC a system that you use for serious work. Ever. Otherwise, yup, the Pentium AE is good bang for the buck.

 

[plopke]

Somebody's getting a time out soon.

Seriously, though. If Zen is 8C/16T, and Sandy Bridge IPC, that's not a bad thing(*). Really, at that point, it comes down to pricing, both unit SKU, and TCO, which involves some performance/watt calculations. (Which I believe Phynaz is involved in something related to that professionally, no?)

(*) For general usage, I mean. Gaming benefits from high-IPC / high-clocked CPUs, and much less from multi-core, at least thus far. (See: games that "require" a quad-core, run fine on a quad-threaded dual-core.)

If this is it, still looks like a solid product maybe not for gaming. They might actually get some of that more profitable enterprise money again , AMD needs profits not gamers approval.

Curious how game design will evolve , if it isn't to expensive and can compete with core i5 we might see for a first time 8 threads becoming standard on all platforms instead of just consoles.


But on a side note din't the ES of RX 480 show up first with a 800Mhz core frequency?

 

[The Stilt]

If you guy search a little you find that AMD's 8350 engineering sample was clocked at 3ghz and final was 4ghz, so there is quite some room for improvement here...

Except that's wasn't the case. The second major prototype stepping (OR-A1, ZD262046W8K43 SKU) already reached 3.6GHz, which is 400MHz lower than the fastest retail part (FX-8120) did at the same TDP.
A1 stepping of Bulldozer was taped out a year (September-October 2010) before the final product was launched. That's why the situation is different to Zeppelin, since obviously we are not talking about a product still year away (hopefully at last).

 

[The Stilt]

Do we know if 4c/8t part will be a cut down die? I guess if they fab it at glofo they will probably save some money on designing 'small die' Zen and use defective or 'artificially defective' dies. Anyone in for unlocking to 8c/16t from 4c/8t a'la old athlon x2 to phenom x4 ? They surely can 'waste' a lot of GF silicon since they will have to pay for unused capacity

I hope they get the APUs right and are not afraid to pump some HBM into the mix. If they release AM4 apu with HBM at 125W or more I'm buying it day 1.

There is a single Zeppelin die. Harvesting will allow AMD to create parts with 2, 4, 6 core SKUs with 0MB, 8MB or 16MB L3 upon demand. Zeppelin is a multipurpose core, same way Orochi was.