• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD Zen 2 (Ryzen 3000) - *** NO COMPETITOR HINTING ***

I've got to be honest, my 2700x has been a disappointment for gaming; it holds back my 1080ti too often with a single thread getting tortured while the rest snooze. I really hope Zen 2 can live up to the hype and hit the 5GHz threshold else I'll be back on Intel. :(
 
I've got to be honest, my 2700x has been a disappointment for gaming; it holds back my 1080ti too often with a single thread getting tortured while the rest snooze. I really hope Zen 2 can live up to the hype and hit the 5GHz threshold else I'll be back on Intel. :(

I’m in the same boat with my 1700 - at 4ghz it struggles in some games that were solid on my 4770k @ 4.3ghz. Even if 5ghz is a single thread boost clock, it would make a huge difference in some scenarios.
 
I’m in the same boat with my 1700 - at 4ghz it struggles in some games that were solid on my 4770k @ 4.3ghz. Even if 5ghz is a single thread boost clock, it would make a huge difference in some scenarios.

Think in situations like yours just the expected IPC gain should make a difference, any increase in ghz is a bonus :D
 
:confused: 5.3GHz at only -40 degrees is a very good result for 1st generation 7nm process, and being "only" -40 degrees it bodes well that at normal operating temperatures a 7nm design could operate at 5GHz, and therefore by extension 7nm chiplets could hit that frequency.

5.3 GHz should be with good water cooling, if 5 GHz goes with top air cooling. This is what I expect. -40°C is a very serious minus temperature. How are you going to get it?
 
5.3 GHz should be with good water cooling, if 5 GHz goes with top air cooling. This is what I expect. -40°C is a very serious minus temperature. How are you going to get it?

That isn't what he is saying at all. He is referring to the latest gen under water cooling against a 1st gen with liquid nitrogen and not being that cold.
 
That isn't what he is saying at all. He is referring to the latest gen under water cooling against a 1st gen with liquid nitrogen and not being that cold.

I don't think so.

His post:

5.3GHz at only -40 degrees is a very good result for 1st generation 7nm process, and being "only" -40 degrees it bodes well that at normal operating temperatures a 7nm design could operate at 5GHz, and therefore by extension 7nm chiplets could hit that frequency.

The truth:

5.3GHz at -40°C degrees is not a good result for 1st generation 7nm process, and being -40°C degrees it doesn't bode well that at normal operating temperatures a 7nm design could operate at 5GHz, and therefore by extension 7nm chiplets could hit that frequency.
 
I don't think so.

His post:



The truth:

5.3GHz at -40°C degrees is not a good result for 1st generation 7nm process, and being -40°C degrees it doesn't bode well that at normal operating temperatures a 7nm design could operate at 5GHz, and therefore by extension 7nm chiplets could hit that frequency.

5.3Ghz at -40c indeed would not bode well at all for normal air or water cooling. As someone that has spent the best part of the last 16 years or so using and then building Phase Change units, that would be a rubbish result. At -40 that would only be an average SS unit in the first place. The one i have at the moment will hold 300 Watts at -50c 24/7 without breaking a sweat. The point is though, the reference to 5.3Ghz at -40c was in a document that is nearly 2 years old, the whole process would have come on in leaps and bounds since then so the whole document is so well outdated as to be worthless.
 
My point was TSMC were able to produce a L1 cache design that operated above 5GHz using their 7nm process, which as Beren points out, was quite early on in its development. This is significant for 2 reason:

1: L1 cache frequency determines CPU max frequency, so if a L1 cache design can operate at 5GHz then the CPU as a whole will not be hampered by cache frequency limits
2: TSMC demonstrated their manufacturing process is capable of reaching those frequencies, so the node itself would not be a performance bottleneck, as was the case with GloFo's 14nm and 12nm.

So what I was saying is that the Zen 2 architecture operating at 5GHz is not an impossibility; both the cache design and the process node are capable of doing so.
 
My point was TSMC were able to produce a L1 cache design that operated above 5GHz using their 7nm process, which as Beren points out, was quite early on in its development. This is significant for 2 reason:

1: L1 cache frequency determines CPU max frequency, so if a L1 cache design can operate at 5GHz then the CPU as a whole will not be hampered by cache frequency limits
2: TSMC demonstrated their manufacturing process is capable of reaching those frequencies, so the node itself would not be a performance bottleneck, as was the case with GloFo's 14nm and 12nm.

So what I was saying is that the Zen 2 architecture operating at 5GHz is not an impossibility; both the cache design and the process node are capable of doing so.

So the architectual design has little bearing on a CPU's speed? So if Ryzen was fabricated at Intel's plant, they would be nailing 5GHz? And if GloFo was given Skylake-X plans, they would be limited to around 4.1GHz on all cores?

So if TSMC really is that good, higher clock speeds are guaranteed. A 15% gain on clocks would be amazing.
 
So the architectual design has little bearing on a CPU's speed? So if Ryzen was fabricated at Intel's plant, they would be nailing 5GHz? And if GloFo was given Skylake-X plans, they would be limited to around 4.1GHz on all cores?

So if TSMC really is that good, higher clock speeds are guaranteed. A 15% gain on clocks would be amazing.

Obviously arch design has a major impact. The dissatisfaction with the GloFlo 14/12nm process has also been very public though.

No-one can say with absolute certainty but it is clear that GloFlo have been a major roadblock for AMD.

Did GloFlo ever publicly show a capability of 5+Ghz like TSMC are showing on their 7nm node?

What we are going to see Soon® is exactly what Zen is capable of. If it can't hit 5ghz then we will know there is a architectural limitation holding it back. If it achieves 4.8+ all core then it will give an indication as to the scope of the issues they have had.
 
So the architectual design has little bearing on a CPU's speed? So if Ryzen was fabricated at Intel's plant, they would be nailing 5GHz? And if GloFo was given Skylake-X plans, they would be limited to around 4.1GHz on all cores?

So if TSMC really is that good, higher clock speeds are guaranteed. A 15% gain on clocks would be amazing.

You're overthinking, but it's entirely possible that the answer to your questions is "yes". It's very public that GloFo's 14nm was **** and it's very public that Intel's 14nm+++++++ is great. So it is entirely possible that if GloFo made a 14nm CPU like Skylake and later then it wouldn't perform anywhere near as well. Similarly, Zen 1 built by Intel's fab could have been much better than it already was.

But all this what if is irrelevant. All I was saying is TSMC 2 years ago demonstrated the capability to get a 5.3GHz frequency out of a L1 cache design on their fledgling 7nm process. That shows TSMC's 7nm node is capable of manufacturing silicon with high frequency potential. In turn that means Zen 2 is unlikely to be hampered by manufacturing shortcomings.

In short, the final numbers we will see with Zen 2 in all of its guises will be representative of the architecture, purely and solely. We'll see what AMD is capable of, no excuses.
 
If these perform really well it's going to be a decision of whether to wait for the chips after, just for future proofing and DDR5.

Though it does look as if Zen 2 will be a leap, and the next one or 2 chips after will be rather small steps.
 
I am sat on the fence on building a Ryzen setup to replace my 4770k, the above comments are making me think that I need to wait for the 3000 series as the 1 & 2 aren't going to make enough difference over what I already have

This thread has been a great read overall massively helpful as my knowledge on AMD these days is limited where as I always swore by them up to the 939 x2 cores in my glory days :D
 
Well First post I read this morning was someone running a 4770k at same clock as me, was thinking a 1700 so made me think about it a bit more

Saved from maybe a bad spend = helpful :D
 
Well First post I read this morning was someone running a 4770k at same clock as me, was thinking a 1700 so made me think about it a bit more

Saved from maybe a bad spend = helpful :D
The other side of that is I have just upgraded from a 4790k at 4.3ghz to a stock ryzen 2600 (stopgap as am swapping out for zen 2), I am totally impressed, I am not one for counting exact FPS but so far on everything I am playing it holds the 75fps (freesync) that I need and in general its a much smoother system multitasking, I appreciate everyone's setup and usage is different but am very happy I made the jump.

Just my 2 Pence.

EDIT - That's with a Vega 56 Pulse and 3200mhz ram for comparison
 
The other side of that is I have just upgraded from a 4790k at 4.3ghz to a stock ryzen 2600 (stopgap as am swapping out for zen 2), I am totally impressed, I am not one for counting exact FPS but so far on everything I am playing it holds the 75fps (freesync) that I need and in general its a much smoother system multitasking, I appreciate everyone's setup and usage is different but am very happy I made the jump.

Just my 2 Pence.

EDIT - That's with a Vega 56 Pulse and 3200mhz ram for comparison
Thanks chap

That's interesting to know, that was my kind of idea too, go for a stop gap CPU and 450 board that (should) accept zen2. Was originally thinking 1700 but just been looking at 1600x, they are in the £120 region!!!
 
Back
Top Bottom