Ryzen Allegedly Matching Intel in 3DMark FS Physics

[stconquest]

7 minutes ago, techstorm970 said:

You're right.  Still, 5960X was king just a year ago.

Yup.

 

The enthusiast line seems to always be a year behind the mainstream/performance line.

[samiscool51]

oh yea i forgot to post this

[Enderman]

19 minutes ago, dizmo said:

Two? Isn't it one?

My bad, it's actually 3.

The new 7000 is kabby lake.

The 6000 series is skylake.

The 6950X is not skylake, but broadwell, same as the 5770C.

The 5960X is devils canyon, same as the 4790K.

Skylake-E is not yet released, it should be called the 7950X and will be 1 gen behind the 7700K.

That means the 6950X is 2 gens behind and the 5960X is 3 gens behind.

[Techstorm970]

1 hour ago, Enderman said:

Based on these benchmarks, I've done a new analysis of the R3 series.  No R3 processors are shown here, but we can probably expect the single and quad core performance of an R3-1100 to be similar to an R5-1300 (which is shown).

 

i5-3570k <= R3-1100 < i5-7500 < R3-1200X <= i5-4690k

[dizmo]

1 hour ago, Enderman said:

My bad, it's actually 3.

The new 7000 is kabby lake.

The 6000 series is skylake.

The 6950X is not skylake, but broadwell, same as the 5770C.

The 5960X is devils canyon, same as the 4790K.

Skylake-E is not yet released, it should be called the 7950X and will be 1 gen behind the 7700K.

That means the 6950X is 2 gens behind and the 5960X is 3 gens behind.

I suppose if you look at it that way.

However, it's only one generation behind if you're talking about the -E nomenclature.

[Enderman]

1 minute ago, dizmo said:

I suppose if you look at it that way.

However, it's only one generation behind if you're talking about the -E nomenclature.

No, I'm talking about the actual generation, based on process, which is why stuff has the names "kaby lake" and whatnot.

That's what determines the core performance.

[typographie]

1 hour ago, dizmo said:

However, it's only one generation behind if you're talking about the -E nomenclature.

You can look at the HEDT product line in a vacuum if you prefer, but that's not how Intel does business. Broadwell is Broadwell. Broadwell-E CPUs have the same IPC as the Broadwell-S CPUs like the 5775C. It's not just "another way to look at it," Kaby Lake is different tech with different features and performance.

[MageTank]

On 2/14/2017 at 8:23 AM, Bouzoo said:

What didn't he do?

Time for @MageTank to make a thread to reverse that decision. Batman can't work without his Joker. 

I have no control over what the mods do. That being said, the mods were very lenient towards both he and myself, even when our banter derailed threads entirely. He had plenty of chances, but refused to change. It's nothing short of a miracle that I've made it this far on this forum without a single warning point. I guess explaining things to people without calling them names tends to look better in the eyes of the mods?

 

@ the topic: If this is indeed true, Zen is looking like a pretty good deal. Especially after seeing that recent price leak, suggesting a $300 1700 Pro. $300 for an unlocked 8c/16t CPU is very substantial, even if it's IPC were to end up worse than Haswell. That's $300 for a CPU that is perfectly capable of heavy lifting on a budget. Even if it were to only match Sandy or Ivy, $300 for that many threads, it's gonna sell well. I know plenty of streamers that currently use i5's and quad core i7's that tell me they still need more cores, and this would be the answer for them. 

 

Hopefully we get reviews soon, so we can know for certain what kinds of IPC we are talking, and what the differences between the $300 and $500 16T SKU's are. I personally don't see these overclocking that well (maybe 4.5ghz at best) but then again, 16 threads of 4.5ghz is also plenty enough for most consumers. 

[Bouzoo]

2 minutes ago, MageTank said:

snip

That my friend, was a joke. 

[MageTank]

1 minute ago, Bouzoo said:

That my friend, was a joke. 

Sorry, long work week and am very tired. Was absent from the forum for 3 days, lol. 

[Energycore]

17 hours ago, Enderman said:

My bad, it's actually 3.

The new 7000 is kabby lake.

The 6000 series is skylake.

The 6950X is not skylake, but broadwell, same as the 5770C.

The 5960X is devils canyon, same as the 4790K.

Skylake-E is not yet released, it should be called the 7950X and will be 1 gen behind the 7700K.

That means the 6950X is 2 gens behind and the 5960X is 3 gens behind.

One has to keep in mind how little gains are enough for Intel to call them a "new gen". From Broadwell to Kaby Lake, only once has IPC gone up, and only twice from Haswell to Kaby. Ever since Kaby brought no per-clock performance increase and only a little clockspeed, it's not entirely accurate to look at a new gen as a new step in performance.

[Inkz]

On 2/13/2017 at 8:45 PM, SamStrecker said:

If this flops AMD will never come back at all.

The amount of agrees on this is insane.

[Enderman]

29 minutes ago, Energycore said:

One has to keep in mind how little gains are enough for Intel to call them a "new gen". From Broadwell to Kaby Lake, only once has IPC gone up, and only twice from Haswell to Kaby. Ever since Kaby brought no per-clock performance increase and only a little clockspeed, it's not entirely accurate to look at a new gen as a new step in performance.

 

[Drak3]

19 hours ago, dizmo said:

Two? Isn't it one?

One includes the generation that is being spoken of in the count of how many generations old it is.

 

Overall, it's been surpassed 3 times (talking just core architecture), making it 4 generations old.

 

On the HEDT side of things, it has only been surpassed once, by Broadwell-E. That makes it 2 generations old, in HEDT regards.

 

As far as Intel's perception of their CPU generations, Haswell-E and Haswell Mini (AKA Broadwell) are both part of the same generation (gen 5, and Intel is currently on gen 7), as they nearly every feature and HEDT lags behind consumer product. In this sense, Haswell-E is 3 generations old.

 

So then we've got Haswell, Haswell Mini, Haswell Mini 2.0 (Skylake), and Haswell Mini 2.1 (Kaby), with little worthwhile improvement in between them, as far as many people are concerned.

[Edgar R. Zakarian]

All benchmarks are on 1700x, is there an 1800x as well, which is even MORE powerful? or is the 1700x top whale ?

[MageTank]

4 minutes ago, Edgar R. Zakarian said:

All benchmarks are on 1700x, is there an 1800x as well, which is even MORE powerful? or is the 1700x top whale ?

There is supposedly an 1800 and 1800X. The 1700X seems to be the middle ground, as there is the 1700 and 1700 Pro that are below it, while the 1800 and 1800X are above it. It's hard to keep up with this though, as every day a new leak changes how many SKU's exist.

[79wjd]

22 hours ago, ace_cheaply said:

Do Z series boards run all cores at turbo boost speed? I remember seeing that somewhere, I believe.

The board doesn't really affect turbo speeds. Turbo speeds are dependent on temps and the number of cores under load. For example, take the i5-6400 and i5-6500, the 6400 turbos at: 3.3/3.3/3.1 while the 6500 turbos at 3.6/3.5/3.3 (1/2/4 cores).

 

https://en.wikipedia.org/wiki/Skylake_(CPU)

https://en.wikipedia.org/wiki/Kaby_Lake_(CPU)

3 hours ago, Drak3 said:

@leadeater @MageTank 

Rejoice!

@leadeater

I'm still sad he got banned. He was supposed to leave with his tail tucked between his legs once Zen actually launched: 

Spoiler

 

 

 

 

[leadeater]

54 minutes ago, djdwosk97 said:

 

@leadeater

I'm still sad he got banned. He was supposed to leave with his tail tucked between his legs once Zen actually launched: 

As you know we can't really talk about moderation or bans of accounts, neither is it supposed to be talked about by other members in public.

 

Personally I don't mind the guy, I like talking to people with intelligence. However he had an extreme lack of social skills, the inability to know when he was wrong and unable to recognize when someone was more experienced and knowledgeable in areas than he was. He was given many reminders to act respectful to other members of the community, correctness or self belief isn't justification to act condescending/arrogant. His attitude towards other members of the community is something I very strongly disliked, that is no secret. However due to the arguments I ended up having with him I decided not to moderate any of his content and leave it to the rest of the team.

 

He was much less knowledgeable in hardware topics than he thought and acted, that AMD/Xeon thing is a prime example. Any self respecting 'expert' would never mistake that for anything other than an AMD CPU. Having been in the IT industry as long as I have, plus the years of personal use of server equipment when I was young (free kit from parents work), along with building hundreds of computers for computer labs in schools I have seen every design of Intel and AMD CPU. It was an AMD design, no question.

 

Now I will ask that this not get brought up again, it is against the CS plus derails threads. Neither do we need to talk about such things anyway, it's not productive and nothing will come of it.

[MageTank]

3 hours ago, djdwosk97 said:

The board doesn't really affect turbo speeds. Turbo speeds are dependent on temps and the number of cores under load. For example, take the i5-6400 and i5-6500, the 6400 turbos at: 3.3/3.3/3.1 while the 6500 turbos at 3.6/3.5/3.3 (1/2/4 cores).

 

https://en.wikipedia.org/wiki/Skylake_(CPU)

https://en.wikipedia.org/wiki/Kaby_Lake_(CPU)

Some SKU's even impose a power limit, that overrides turbo boost. Even if only one core is boosting, and even if thermals are perfectly in line, it will still throttle the turbo to remain within a specific power envelope. The T SKU's do this, and some of the mobile SKU's do this depending on how the board manufacturers implement it. My brothers XPS13 will impose a power limit, even after getting thermals under control. At best, we are able to change the duration of the boost before the limit kicks in, but only by a few seconds.

 

My worry is, if these 8c/16t SKU's indeed have a 65w TDP, something has to be limiting them in some regard. I simply don't believe they are still boosting to 3.7ghz on 8c/16 threads and are still cutting Intel's TDP directly in half. Some sort of sorcery is at play here.

 

EDIT: I forgot to mention, boards do impact turbo speed. For example: Asrock boards have an "Advanced Turbo" sub-section in their BIOS that allows you to run at your maximum defined boost multiplier on all cores. This is available on all of their Z series boards, and a few of their "gaming" H series boards. I have yet to see it on the H110 or B150 boards that I've tested. That being said, it won't impact how high you can actually boost, only the amount of cores that can simultaneously boost. It ignores the factory defined boost table.

[79wjd]

3 minutes ago, MageTank said:

Some SKU's even impose a power limit, that overrides turbo boost. Even if only one core is boosting, and even if thermals are perfectly in line, it will still throttle the turbo to remain within a specific power envelope. The T SKU's do this, and some of the mobile SKU's do this depending on how the board manufacturers implement it. My brothers XPS13 will impose a power limit, even after getting thermals under control. At best, we are able to change the duration of the boost before the limit kicks in, but only by a few seconds.

 

My worry is, if these 8c/16t SKU's indeed have a 65w TDP, something has to be limiting them in some regard. I simply don't believe they are still boosting to 3.7ghz on 8 threads and are still cutting Intel's TDP directly in half. Some sort of sorcery is at play here.

Good point. 

 

I think the bigger point of "concern" is the large price difference between the 65w 1700x and the 95w 1800x. It's possible AMD is just selling the 1800x for noticeably more because it's clocked higher and other than that they're identical (and can be overclocked to the same degree), but I think that's unlikely. I'm assuming there is a power limit in place (or far less likely -- imo -- less cache). 

[Prysin]

3 hours ago, djdwosk97 said:

The board doesn't really affect turbo speeds. Turbo speeds are dependent on temps and the number of cores under load. For example, take the i5-6400 and i5-6500, the 6400 turbos at: 3.3/3.3/3.1 while the 6500 turbos at 3.6/3.5/3.3 (1/2/4 cores).

 

https://en.wikipedia.org/wiki/Skylake_(CPU)

https://en.wikipedia.org/wiki/Kaby_Lake_(CPU)

@leadeater

I'm still sad he got banned. He was supposed to leave with his tail tucked between his legs once Zen actually launched: 

  Reveal hidden contents

  Reveal hidden contents

 

 

 

actually. Some boards BIOS settings chanes turbo behavior. Its been that way since Ivy Bridge AFAIK. 

My Z97 HERO board allow my 4790k to run 4.4GHz on all 4 cores, despite that Intels official speed for 4 core turbo is something like 4.1 GHz... Lower end boards, like many H87 and H97 boards does not allow for this behavior.

[MageTank]

1 minute ago, Prysin said:

actually. Some boards BIOS settings chanes turbo behavior. Its been that way since Ivy Bridge AFAIK. 

My Z97 HERO board allow my 4790k to run 4.4GHz on all 4 cores, despite that Intels official speed for 4 core turbo is something like 4.1 GHz... Lower end boards, like many H87 and H97 boards does not allow for this behavior.

Yeah, plenty of manufacturers have advanced turbo options to ignore boost tables. I've also yet to see anyone complain about stability when using these features, so it's pretty good. Some higher end laptops have the same thing, though they mostly require some sort of software or "gaming mode" enabled. Worked on an ROG laptop that had such a feature. 

[Prysin]

4 minutes ago, MageTank said:

Yeah, plenty of manufacturers have advanced turbo options to ignore boost tables. I've also yet to see anyone complain about stability when using these features, so it's pretty good. Some higher end laptops have the same thing, though they mostly require some sort of software or "gaming mode" enabled. Worked on an ROG laptop that had such a feature. 

for most Z series ROG boards, it is the default behavior. You cannot turn it on or off without fully disabling Turbo speed. It doesnt really seem to jack up voltages uneccessarily high either. Like "gaming mode"or "Ez OC" stuff usually does.

[MageTank]

1 minute ago, Prysin said:

for most Z series ROG boards, it is the default behavior. You cannot turn it on or off without fully disabling Turbo speed. It doesnt really seem to jack up voltages uneccessarily high either. Like "gaming mode"or "Ez OC" stuff usually does.

We can turn it off on our Asrock boards, but it's not like it really matters. You buy a Z series board to overclock, and overclocking overrides boost anyways, lol. 

[Prysin]

3 minutes ago, MageTank said:

We can turn it off on our Asrock boards, but it's not like it really matters. You buy a Z series board to overclock, and overclocking overrides boost anyways, lol. 

I wonder how it will be with RYZEN boards. The feature is prob going to make it into X370 boards. But how will the voltage and turbo work given RYZEN basically have "CPU boost 3.0"