We are talking about software supposedly written to take advantage of an 8 core CPU, weak or not.
The software should already be written for 8 cores, so if that were any advantage on the desktop, we should have seen it long ago.
The architecture of the 8 core CPUs in the Xbox/PS doesn't change anything at all as far as the software being written for multiple cores. They are still 8 core CPUs.
I never heard of 2 cores being reserved for the OS. How do you do that?Actually in both consoles one (or was it two?) of the cores are reserved for the OS anyways. And you also can't expect developers to be able to perfectly thread out a game, there will be limiting factors. Nevertheless, 6 core cpus do have advantages over 4 these days. I think it's in line with console core counts, considering other factors.
Rumours say there will be roughly a 10% IPC uplift which would give AMD a parity in games.
Lisa Su's career won't depend on beating the i9-9900k in gaming benchmarks, and I say that as someone who happily upgraded to a 2700 last week. They're coming from both an IPC and clockspeed deficit. There's no guarantee they'll surpass Intel's mainstream halo product in a workload it's historically done better in. AMD doesn't need to smash Intel into the dust, they need to offer a compelling product and so far they've done that.Do you simply think she's just overly confident and will fall short?
A lot of "arguments" are faith based. We'll see soon enough.As posted in another thread, today's rumour from Bits-And-Chips is 13% in scientific workloads:
"Zen+ -> Zen2: +13% IPC (Average) in scientific tasks. Not bad. P.S. No gaming data, atm."
https://mobile.twitter.com/BitsAndChipsEng/status/1052194745647165441
How much was i9-9900K ahead of the Ryzen 2700X in the revised PT study again? I've seen various numbers quoted. But let's say 16%, including outliers such as the primordial CS:GO title. Assuming Ryzen 3000 has 13% IPC uplift and a mere 10% frequency uplift, we're at 24% uplift total (1.13 * 1.10 = 1.243), putting it well ahead of i9-9900K — more so, if you eliminate outliers.
This sounds reasonable. So why do so many here expect so little of AMD? 50% of those voting so far trust Intel to retain or regain the lead, and of those of you who voted "no" and have put your thoughts into a post, most of you doubt Zen 2 will even catch up with i9-9900K. I thought most of the doubters would concede that Zen 2 would probably leap ahead, but that Intel would come back next autumn with a rabbit out of the hat. But I am surprised to see that AMD mindshare is even lower than that.
Come on! Let's expect something from AMD! Not even matching a 14nm Skylake-derivative by 2019 would be a resounding failure of planning and execution. If so, AMD would be doomed hadn't Intel stumbled with 10nm. Clearly, it would have been a downright folly of AMD CEO Lisa Su to embark on head-to-head competition with Intel in high-performance x86, if she thought they could only target Skylake performance by 2018-2019, when Intel roadmaps showed 10nm and Ice Lake in the same timeframe. That does not make any sense. She must have been confident about competing against a well-performing 10nm Ice Lake by now, for her to decide to squarely focus on high-performance x86 again.
Note that her laser focus included dropping (or parking) many projects, including K12, Jim Keller's custom ARM core, along with Rory Read's entire ambidextrous strategy with ARM, which was devised to shield AMD somewhat from direct Intel competition by spearheading a new niche — Lisa Su threw all that out and said "brush up and let's go fight Intel again at the x86 high end".
Do you simply think she's just overly confident and will fall short?
“Rome was designed to compete favorably with “Ice Lake” Xeons.
– Forrest Norrod.
That article was just an AMD ad, though.https://www.nextplatform.com/2018/06/20/amds-epyc-return-to-the-datacenter-ring/
Seems that Zen 2 shouldnt have much trouble competing against CFL, more than favorably...
As posted in another thread, today's rumour from Bits-And-Chips is 13% in scientific workloads:
"Zen+ -> Zen2: +13% IPC (Average) in scientific tasks. Not bad. P.S. No gaming data, atm."
https://mobile.twitter.com/BitsAndChipsEng/status/1052194745647165441
How much was i9-9900K ahead of the Ryzen 2700X in the revised PT study again? I've seen various numbers quoted. But let's say 16%, including outliers such as the primordial CS:GO title. Assuming Ryzen 3000 has 13% IPC uplift and a mere 10% frequency uplift, we're at 24% uplift total (1.13 * 1.10 = 1.243), putting it well ahead of i9-9900K — more so, if you eliminate outliers.
This sounds reasonable. So why do so many here expect so little of AMD? 50% of those voting so far trust Intel to retain or regain the lead, and of those of you who voted "no" and have put your thoughts into a post, most of you doubt Zen 2 will even catch up with i9-9900K. I thought most of the doubters would concede that Zen 2 would probably leap ahead, but that Intel would come back next autumn with a rabbit out of the hat. But I am surprised to see that AMD mindshare is even lower than that.
Come on! Let's expect something from AMD! Not even matching a 14nm Skylake-derivative by 2019 would be a resounding failure of planning and execution. If so, AMD would be doomed hadn't Intel stumbled with 10nm. Clearly, it would have been a downright folly of AMD CEO Lisa Su to embark on head-to-head competition with Intel in high-performance x86, if she thought they could only target Skylake performance by 2018-2019, when Intel roadmaps showed 10nm and Ice Lake in the same timeframe. That does not make any sense. She must have been confident about competing against a well-performing 10nm Ice Lake by now, for her to decide to squarely focus on high-performance x86 again.
Note that her laser focus included dropping (or parking) many projects, including K12, Jim Keller's custom ARM core, along with Rory Read's entire ambidextrous strategy with ARM, which was devised to shield AMD somewhat from direct Intel competition by spearheading a new niche — Lisa Su threw all that out and said "brush up and let's go fight Intel again at the x86 high end".
Do you simply think she's just overly confident and will fall short?
That article was just an AMD ad, though.
What about the substance, i mean, do you have any argument and info that point this AMD rep as being inaccurate, that is, that they designed Zen 2 with SKL, instead of ICL, as target..?..
Zen+ perf/clock improvement was 3% for IPC and 5% for games, we can extrapolate with this basis for Zen+ vs Zen 2.
If the numbers that float around are on point then we can expect that said ~13% better perf/clock in applications to be enough for something like 20% better perfs/clock in games, this would be in line with Lisa Su s statement that Zen 2 would be "much better" for games...
When I see "sponsored content", I don't see any point in continuing. When I look at other AMD articles from the guy and see "sponsored content" again, I give up.What about the substance, i mean, do you have any argument and info that point this AMD rep as being inaccurate, that is, that they designed Zen 2 with SKL, instead of ICL, as target..?..
Zen+ perf/clock improvement was 3% for IPC and 5% for games, we can extrapolate with this basis for Zen+ vs Zen 2.
If the numbers that float around are on point then we can expect that said ~13% better perf/clock in applications to be enough for something like 20% better perfs/clock in games, this would be in line with Lisa Su s statement that Zen 2 would be "much better" for games...
Maybe with good cooling? I doubt though that Intel's 10nm will hit 5Ghz any time soon.Will Zen2 really hit 5ghz though? That's really Intel's big advantage right now. I see a bunch of people in this thread claiming it's a foregone conclusion but is there data to support that conclusion?
Maybe with good cooling? I doubt though that Intel's 10nm will hit 5Ghz any time soon.
Its common knowledge that 1 CPU core is preserved for the OS, Sony had 2 but enabled a seventh core through a patch.I never heard of 2 cores being reserved for the OS. How do you do that?
Here is a blurb on what the XBOX has in it currently:
https://www.anandtech.com/show/11992/the-xbox-one-x-review/3
Let's not forget unprecedented problems surfacing when designing chips for 7nm. Problems magnified horribly by additional voltage (required for super-high clock speeds)My opinion as to why Zen 2 won't hit 5GHz:
1. Fresh TSMC 7nm process.
2. Precedent. Has a TSMC process ever hit 5GHz before? Ever?
3. Is AMD going to be able to successfully marry the zen 2 arch with this fresh TSMC process to produce a 5GHz-capable chip out of the gates?
4. A process shrink means smaller cooling surface area. High frequency and voltage will mean concentration of heat in a reduced core area which will need to be moved as quickly as possible. This will pose its own challenges.
5. Is there any other process out there, besides Intel's, that's consistently hitting 5GHz on air cooling?
No idea. Never heard it before. Haven't used a console in a very long time.Its common knowledge that 1 CPU core is preserved for the OS, Sony had 2 but enabled a seventh core through a patch.
You really think they wouldn't do this to enable seamless transitions between the OS/game and streaming?
https://www.extremetech.com/gaming/...-ps4s-seventh-cpu-core-but-will-games-benefitNo idea. Never heard it before. Haven't used a console in a very long time.
That seems like a strange way to run a railroad.
Will Zen2 really hit 5ghz though?
My opinion as to why Zen 2 won't hit 5GHz:
1. Fresh TSMC 7nm process.
+1. All excellent points. It's possible AMD can hit 5Ghz but I doubt they will for those same reasons. They'd break the Internet if they pulled it off, though.My opinion as to why Zen 2 won't hit 5GHz:
1. Fresh TSMC 7nm process.
2. Precedent. Has a TSMC process ever hit 5GHz before? Ever?
3. Is AMD going to be able to successfully marry the zen 2 arch with this fresh TSMC process to produce a 5GHz-capable chip out of the gates?
4. A process shrink means smaller cooling surface area. High frequency and voltage will mean concentration of heat in a reduced core area which will need to be moved as quickly as possible. This will pose its own challenges.
5. Is there any other process out there, besides Intel's, that's consistently hitting 5GHz on air cooling?