Discussion: AMD Ryzen

[8350rocks]

http://www.anandtech.com/bench/CPU/1028

Even if you compare the 6350 and the i3-4360, the 6350 is 3.6 stock, and the i3 is 3.7 stock

FX6350 is 96 in ST, and i3-4360 is 149 in ST.

40% is 134

50% is 144

55% is 148

Hmm...148 vs 149, with intel having a negligible clock advantage, or just about haswell. That is going from PD as well...not Excavator.

The FX-6350 has a base clock of 3.9GHz, not 3.6. Moreover, this is ST and one has to use turbo clocks instead. The FX-6350 has a single-core turbo of 4.2GHz. This is the reason why it gets the same ST score than the FX-8350 (also 4.2GHz turbo) which appears in the list with 96 points as well. The i3-4360 doesn't have turbo and it runs all the benchmark at 3.7GHz. Correcting for the clocks we obtain

(149 / 96) * (4.2 / 3.7) = 1.76

I.e. Haswell IPC is about 76% above Piledriver, not 55%. The same conclusion can be obtained from the 3GHz data I provided above

(121 / 68.10) = 1.75

One can also see that the presence or absence of L3 doesn't alter the result in any significant way: 76% (using a PD CPU with L3) versus 75% (using a PD APU without L3). This puts to a rest the argument that I was using an incorrect baseline for my early estimations of Zen IPC, when I was using a real Excavator-based Athlon CPU (no L3) as baseline, instead using a non-existent Excavator-based FX CPU with L3.

The benchmarks show the 6300 @ 3.6, I mistakenly put 6350. Check the benchmarks, they are clearly labeled and are from Anandtech.

 

[8350rocks]

http://www.anandtech.com/bench/CPU/1028

Even if you compare the 6350 and the i3-4360, the 6350 is 3.6 stock, and the i3 is 3.7 stock

FX6350 is 96 in ST, and i3-4360 is 149 in ST.

40% is 134

50% is 144

55% is 148

Hmm...148 vs 149, with intel having a negligible clock advantage, or just about haswell. That is going from PD as well...not Excavator.

So, I would say, haswell has to be *really* close to what the ST performance will be.

EDIT: Just above the i3 is an i7-6850K (BWE) with a ST score of 152 at same clocks as FX6350, which is right in the ballpark. I think they are really close on this one.

The issue is that no benchmark is universal and not even cinebench R15 in the disclaimer say that they measure true IPC but give only a CPU design throughput. IPC and throughput is not the same.

They measure ST throughput though with that aspect of the benchmark. Those are the single thread benchmarks, multi-thread would clearly have the 6850K significantly higher than both the i3 and the FX 6350.

You didn't factor in turbo, the FX6350 runs a turbo around 4.2 so to score at 3.7 it gets 86. The Bulldozer uarch was fundamentally bad for IPC, even a Deneb and Thuban ran circles around a FX8370 clock for clock.

The point of contention is, if AMD means that their CPU can universally at least increase instructions per cycle by 40, that could alone equate to 50% in one bench while 80-90% in another benchmark. You cannot blanket a generic 40% across every bench.

COMMENT=Results Disclaimer - CINEBENCH results are indicative of overall system performance when using CINEMA 4D,
COMMENT=and do not necessarily reflect the performance of the tested hardware with other applications.
COMMENT=Performance of each component (processor, graphics card) does rely somewhat on other components in the system.
COMMENT=Results provided are typical, although not derived from specific testing procedures.
COMMENT=

Looking back, this was the 6300, so turbo core is actually 3.9 in that benchmark.

 

[8350rocks]

Also as AMD mentioned their goal is power.

AMD also mentioned high-frequency desing. And clustering is a standard technique to achieve higher clocks.

"AMD says the much-touted "40 percent higher IPC" over the Excavator core came from three design goals: core, cache, and power. For the core, AMD made everything bigger and wider, introducing a micro-op cache (something Intel has been using for some time), as well as a larger dispatch, larger retire, larger schedulers, and better branch prediction. On the cache side there's a faster prefetch, while L1 and L2 bandwidth has been doubled, and L3 more then quadrupled. Full details on the improvements are in the slides below."
http://arstechnica.com/gadgets/2016/12/amd-zen-performance-details-release-date/

I know. It doesn't change a bit what we are saying about how distributed schedulers can affect IPC and SMT.

I.e. Haswell IPC is about 76% above Piledriver, not 55%. The same conclusion can be obtained from the 3GHz data I provided above

yes. but you forgot to add the 10-12% improvement EV has over piledriver.

1.55*1.12=1.74 ~ Haswell

No. He was comparing Haswell vs Piledriver, not Haswell vs Excavator. He divided the scores of the FX-6350 and the i3-4360 without considering the clocks and got a 55%

(149 / 96) = 1.55

When we correct for the clocks (the FX-6350 runs that ST bench at 4.2GHz) we obtain 76% instead

(149 / 96) * (4.2 / 3.7) = 1.76

No...it was the 6300 @ 3.6 per the site.

Even if you factor turbo core, consider that 3.9 turbo core 6300 versus 4.1 turbo core for i3-4360 =

149/96*3.9/4.1=1.476

by your own math.

EDIT: Actually...that is off a bit.

Should be:

96/149*3.9/4.1=0.612 or a 38.8% deficit.

That keeps the terms in proper order

 

[juanrga]

http://www.anandtech.com/bench/CPU/1028

Even if you compare the 6350 and the i3-4360, the 6350 is 3.6 stock, and the i3 is 3.7 stock

FX6350 is 96 in ST, and i3-4360 is 149 in ST.

40% is 134

50% is 144

55% is 148

Hmm...148 vs 149, with intel having a negligible clock advantage, or just about haswell. That is going from PD as well...not Excavator.

The FX-6350 has a base clock of 3.9GHz, not 3.6. Moreover, this is ST and one has to use turbo clocks instead. The FX-6350 has a single-core turbo of 4.2GHz. This is the reason why it gets the same ST score than the FX-8350 (also 4.2GHz turbo) which appears in the list with 96 points as well. The i3-4360 doesn't have turbo and it runs all the benchmark at 3.7GHz. Correcting for the clocks we obtain

(149 / 96) * (4.2 / 3.7) = 1.76

I.e. Haswell IPC is about 76% above Piledriver, not 55%. The same conclusion can be obtained from the 3GHz data I provided above

(121 / 68.10) = 1.75

One can also see that the presence or absence of L3 doesn't alter the result in any significant way: 76% (using a PD CPU with L3) versus 75% (using a PD APU without L3). This puts to a rest the argument that I was using an incorrect baseline for my early estimations of Zen IPC, when I was using a real Excavator-based Athlon CPU (no L3) as baseline, instead using a non-existent Excavator-based FX CPU with L3.

The benchmarks show the 6300 @ 3.6, I mistakenly put 6350. Check the benchmarks, they are clearly labeled and are from Anandtech.

No. The FX-6300 doesn't even appear in that list... not to mention that the FX-6300 has a turbo of 4.1GHz.

As everyone can see the score of 96 that you used is the score for a FX-6350, and this is a Piledriver chip with 4.2GHz turbo. Haswell IPC is about 75% ahead of Piledriver.

 

[sarinaide]

Maybe AMD meant 40% faster than the previous best CPU notwithstanding clockspeed which Ryzen is and despite giving up over a Ghz in clockspeed to the 8370 in the CPC benchmark, if you factor clockspeed in which is a 26% difference then maybe overall IPC is 60%+

 

[iceclock]

any idea of performance of 6core Ryzen?

and pricing ty

 

[8350rocks]

http://www.anandtech.com/bench/CPU/1028

Even if you compare the 6350 and the i3-4360, the 6350 is 3.6 stock, and the i3 is 3.7 stock

FX6350 is 96 in ST, and i3-4360 is 149 in ST.

40% is 134

50% is 144

55% is 148

Hmm...148 vs 149, with intel having a negligible clock advantage, or just about haswell. That is going from PD as well...not Excavator.

The FX-6350 has a base clock of 3.9GHz, not 3.6. Moreover, this is ST and one has to use turbo clocks instead. The FX-6350 has a single-core turbo of 4.2GHz. This is the reason why it gets the same ST score than the FX-8350 (also 4.2GHz turbo) which appears in the list with 96 points as well. The i3-4360 doesn't have turbo and it runs all the benchmark at 3.7GHz. Correcting for the clocks we obtain

(149 / 96) * (4.2 / 3.7) = 1.76

I.e. Haswell IPC is about 76% above Piledriver, not 55%. The same conclusion can be obtained from the 3GHz data I provided above

(121 / 68.10) = 1.75

One can also see that the presence or absence of L3 doesn't alter the result in any significant way: 76% (using a PD CPU with L3) versus 75% (using a PD APU without L3). This puts to a rest the argument that I was using an incorrect baseline for my early estimations of Zen IPC, when I was using a real Excavator-based Athlon CPU (no L3) as baseline, instead using a non-existent Excavator-based FX CPU with L3.

The benchmarks show the 6300 @ 3.6, I mistakenly put 6350. Check the benchmarks, they are clearly labeled and are from Anandtech.

No. The FX-6300 doesn't even appear in that list... not to mention that the FX-6300 has a turbo of 4.1GHz.

As everyone can see the score of 96 that you used is the score for a FX-6350, and this is a Piledriver chip with 4.2GHz turbo. Haswell IPC is about 75% ahead of Piledriver.

Why is it listed @ 3.6 GHz if it is the 6350?

EDIT: Also, AMD Turbo Core 3.0 is only 300 MHz, except on FX8350/8370 where it is 400 MHz. Even the 9590 is only a 300 MHz clock boost. No AMD DT chip has a 500 MHz gap between base clock and turbo clock.

 

[salgado18]

any idea of performance of 6core Ryzen?

and pricing ty

No 6 core, only 4 and 8 core, with and without SMT ("hyperthreading")

Princing is heavy speculation right now.

 

[juanrga]

http://www.anandtech.com/bench/CPU/1028

Even if you compare the 6350 and the i3-4360, the 6350 is 3.6 stock, and the i3 is 3.7 stock

FX6350 is 96 in ST, and i3-4360 is 149 in ST.

40% is 134

50% is 144

55% is 148

Hmm...148 vs 149, with intel having a negligible clock advantage, or just about haswell. That is going from PD as well...not Excavator.

The FX-6350 has a base clock of 3.9GHz, not 3.6. Moreover, this is ST and one has to use turbo clocks instead. The FX-6350 has a single-core turbo of 4.2GHz. This is the reason why it gets the same ST score than the FX-8350 (also 4.2GHz turbo) which appears in the list with 96 points as well. The i3-4360 doesn't have turbo and it runs all the benchmark at 3.7GHz. Correcting for the clocks we obtain

(149 / 96) * (4.2 / 3.7) = 1.76

I.e. Haswell IPC is about 76% above Piledriver, not 55%. The same conclusion can be obtained from the 3GHz data I provided above

(121 / 68.10) = 1.75

One can also see that the presence or absence of L3 doesn't alter the result in any significant way: 76% (using a PD CPU with L3) versus 75% (using a PD APU without L3). This puts to a rest the argument that I was using an incorrect baseline for my early estimations of Zen IPC, when I was using a real Excavator-based Athlon CPU (no L3) as baseline, instead using a non-existent Excavator-based FX CPU with L3.

The benchmarks show the 6300 @ 3.6, I mistakenly put 6350. Check the benchmarks, they are clearly labeled and are from Anandtech.

No. The FX-6300 doesn't even appear in that list... not to mention that the FX-6300 has a turbo of 4.1GHz.

As everyone can see the score of 96 that you used is the score for a FX-6350, and this is a Piledriver chip with 4.2GHz turbo. Haswell IPC is about 75% ahead of Piledriver.

Why is it listed @ 3.6 GHz if it is the 6350?

EDIT: Also, AMD Turbo Core 3.0 is only 300 MHz, except on FX8350/8370 where it is 400 MHz. Even the 9590 is only a 300 MHz clock boost. No AMD DT chip has a 500 MHz gap between base clock and turbo clock.

The 3.6GHz is obviously a typo because, as said before, the FX-6350 has a base clock of 3.9GHz. But, once again, the base clock is irrelevant because on a ST bench the chip will run on single-core turbo mode. The FX-6350 has a turbo clock of 4.2GHz. The gap is 300MHz. This is all well-known,

http://www.cpu-world.com/CPUs/Bulldozer/AMD-FX-Series%20FX-6350.html
http://products.amd.com/en-us/search/CPU/AMD-FX-Series/AMD-FX-6-Core-Black-Edition/FX-6350/96

The FX-8350 doesn't have 400MHz gap. The base clock is 4GHz, the turbo is 4.2GHz.
The FX-8370 doesn't have 400MHz gap. The base clock is 4GHz, the turbo is 4.3GHz.

http://www.cpu-world.com/CPUs/Bulldozer/AMD-FX-Series%20FX-8350.html
http://products.amd.com/en-us/search/CPU/AMD-FX-Series/AMD-FX-8-Core-Black-Edition/FX-8350/92

http://www.cpu-world.com/CPUs/Bulldozer/AMD-FX-Series%20FX-8370.html
http://products.amd.com/en-us/search/CPU/AMD-FX-Series/AMD-FX-8-Core-Black-Edition/FX-8370/100

As mentioned before the FX-6350 has the same turbo than the FX-8350, both have turbo of 4.2GHz, this is why both chips score 96 points in Anandtech database.

 

[sarwar_r87]

Haswell IPC is about 76% above Piledriver.

indeed it is. given the number of times you mentioned this, I feel like the you are tyrrng to make a point here.

but, and I feel like a broken record when I say this:
https://www.youtube.com/watch?v=sT1fEohOOQ0
7:48--- Our L3 bandwidth has actually gone 5x compared to previous generation.

XV APUs have no L3, so how do you reckon going from 0 MB L3 to 8 MB L3 constitutes as 5 x bandwidth? i.e. if L3 is 0 MB, bandwidth is also 0

XV based APU is 12% over PD based APUs. zen is 55 % over XV.

hence, zen_ipc = PD_IPC * XV_boost * Zen_boost = 1.55*1.12 * PD_IPC = 1.74 * PD_IPC.

UPDATE:
But I will give you this, I think AMD's PR is doing something very fishy with Zen. they did it with bulldozer and Polaris. amd I feel the coming months may well be a Déjà vu.

 

[salgado18]

indeed it is. given the number of times you mentioned this, I feel like the you are tyrrng to make a point here.

but, and I feel like a broken record when I say this:
https://www.youtube.com/watch?v=sT1fEohOOQ0
7:48--- Our L3 bandwidth has actually gone 5x compared to previous generation.

XV APUs have no L3, so how do you reckon going from 0 MB L3 to 8 MB L3 constitutes as 5 x bandwidth? i.e. if L3 is 0 MB, bandwidth is also 0

XV based APU is 12% over PD based APUs. zen is 55 % over XV.

hence, zen_ipc = PD_IPC * XV_boost * Zen_boost = 1.55*1.12 * PD_IPC = 1.74 * PD_IPC.

UPDATE:
But I will give you this, I think AMD's PR is doing something very fishy with Zen. they did it with bulldozer and Polaris. amd I feel the coming months may well be a Déjà vu.

90% chance they are refering to original Bulldozer, just to get a high enough number. Also, they would compare to the last L3 they made, wouldn't they? Which would be Steamroller. They don't say specific generation names to let people speculate, give a big number, but no clear answer while being true.

 

[sarwar_r87]

90% chance they are refering to original Bulldozer, just to get a high enough number. Also, they would compare to the last L3 they made, wouldn't they? Which would be Steamroller. They don't say specific generation names to let people speculate, give a big number, but no clear answer while being true.

I donot think its BD/SR. for one, going back so far would raise questions. secondly benchmarks do not add up.

they did specify Excavator are baseline for comparison. its just no one bothered to ask them how they get 5x L3 BW when there is no L3 on commercially available XV.

SPECULATION: I think its some prototype CPU (not APU) that they played that had L3 with and decided to go with zen as zen was almost ready.

 

[8350rocks]

http://www.anandtech.com/bench/CPU/1028

Even if you compare the 6350 and the i3-4360, the 6350 is 3.6 stock, and the i3 is 3.7 stock

FX6350 is 96 in ST, and i3-4360 is 149 in ST.

40% is 134

50% is 144

55% is 148

Hmm...148 vs 149, with intel having a negligible clock advantage, or just about haswell. That is going from PD as well...not Excavator.

The FX-6350 has a base clock of 3.9GHz, not 3.6. Moreover, this is ST and one has to use turbo clocks instead. The FX-6350 has a single-core turbo of 4.2GHz. This is the reason why it gets the same ST score than the FX-8350 (also 4.2GHz turbo) which appears in the list with 96 points as well. The i3-4360 doesn't have turbo and it runs all the benchmark at 3.7GHz. Correcting for the clocks we obtain

(149 / 96) * (4.2 / 3.7) = 1.76

I.e. Haswell IPC is about 76% above Piledriver, not 55%. The same conclusion can be obtained from the 3GHz data I provided above

(121 / 68.10) = 1.75

One can also see that the presence or absence of L3 doesn't alter the result in any significant way: 76% (using a PD CPU with L3) versus 75% (using a PD APU without L3). This puts to a rest the argument that I was using an incorrect baseline for my early estimations of Zen IPC, when I was using a real Excavator-based Athlon CPU (no L3) as baseline, instead using a non-existent Excavator-based FX CPU with L3.

The benchmarks show the 6300 @ 3.6, I mistakenly put 6350. Check the benchmarks, they are clearly labeled and are from Anandtech.

No. The FX-6300 doesn't even appear in that list... not to mention that the FX-6300 has a turbo of 4.1GHz.

As everyone can see the score of 96 that you used is the score for a FX-6350, and this is a Piledriver chip with 4.2GHz turbo. Haswell IPC is about 75% ahead of Piledriver.

Why is it listed @ 3.6 GHz if it is the 6350?

EDIT: Also, AMD Turbo Core 3.0 is only 300 MHz, except on FX8350/8370 where it is 400 MHz. Even the 9590 is only a 300 MHz clock boost. No AMD DT chip has a 500 MHz gap between base clock and turbo clock.

The 3.6GHz is obviously a typo because, as said before, the FX-6350 has a base clock of 3.9GHz. But, once again, the base clock is irrelevant because on a ST bench the chip will run on single-core turbo mode. The FX-6350 has a turbo clock of 4.2GHz. The gap is 300MHz. This is all well-known,

http://www.cpu-world.com/CPUs/Bulldozer/AMD-FX-Series%20FX-6350.html
http://products.amd.com/en-us/search/CPU/AMD-FX-Series/AMD-FX-6-Core-Black-Edition/FX-6350/96

The FX-8350 doesn't have 400MHz gap. The base clock is 4GHz, the turbo is 4.2GHz.
The FX-8370 doesn't have 400MHz gap. The base clock is 4GHz, the turbo is 4.3GHz.

http://www.cpu-world.com/CPUs/Bulldozer/AMD-FX-Series%20FX-8350.html
http://products.amd.com/en-us/search/CPU/AMD-FX-Series/AMD-FX-8-Core-Black-Edition/FX-8350/92

http://www.cpu-world.com/CPUs/Bulldozer/AMD-FX-Series%20FX-8370.html
http://products.amd.com/en-us/search/CPU/AMD-FX-Series/AMD-FX-8-Core-Black-Edition/FX-8370/100

As mentioned before the FX-6350 has the same turbo than the FX-8350, both have turbo of 4.2GHz, this is why both chips score 96 points in Anandtech database.

As I said, the 6300 has 3.9 turbo clock, not 4.1.

Either way...SB is still not 50% ahead of PD.

 

[8350rocks]

90% chance they are refering to original Bulldozer, just to get a high enough number. Also, they would compare to the last L3 they made, wouldn't they? Which would be Steamroller. They don't say specific generation names to let people speculate, give a big number, but no clear answer while being true.

I donot think its BD/SR. for one, going back so far would raise questions. secondly benchmarks do not add up.

they did specify Excavator are baseline for comparison. its just no one bothered to ask them how they get 5x L3 BW when there is no L3 on commercially available XV.

SPECULATION: I think its some prototype CPU (not APU) that they played that had L3 with and decided to go with zen as zen was almost ready.

Could have been a desktop ES...I suppose?

 

[sarwar_r87]

As I said, the 6300 has 3.9 turbo clock, not 4.1.

my stock 6300 does 4.1 GHz. but on single core only. all core turbo is 3.9 GHz.
having said that, anandtech benchmarks are at fixed clocks that is listed. otherwise their benchmarks are meaning less.

for example,
Intel (Skylake) Core i7 6700K (91W, $350) at 4.8 GHz scores 206 is CB 15 ST. There is no 4.8Ghz of 6700k, as we all know.

Could have been a desktop ES...I suppose?

yes,
a Athlon 845 that Jaun uses all the time has lower performance than a Phenom ii x4 975 and A6-3650 on FryRender, which is similar to Blender, i.e. there is no way a mere 40% IPC jump will let them achieve BWE performance even if AMD has a monstrous gain from SMT

proof: a haswell i3 is about 120% faster than AMD Athlon 845 in Blender. since zen 8c16t performs roughly 4 % better compared to haswell 8c16 in Blender, we can say a theoretical 2c4t zen will perform like a haswell i3 2c4t. meaning zen is 125% faster compared to Athlon 845 when IPC + SMT is considered in Blender. now if we consider a 50% IPC gain, that will need a SMT gain of 83%!!!!!

http://www.guru3d.com/articles_pages/amd_athlon_x4_845_fm2_review,7.html
http://techfrag.com/2016/04/11/amd-athlon-x-845-benchmarks-surface-beats-intel-pentium-g4400/

conclusion is one of the three is true:
1. SMT gain of 83% while IPC gain of 50%
2. IPC gain of more than 90% while SMT is 40% (SMT is still more than intel)
3. Athlon 845 is not the right reference model

I think its no 3. Jaun?

EDIT: I compared Athlon 845 with a 2c4t zen because
1. both have same no of FP,
2. 8c16 intel = 8c16t zen. meaning i3 = 2c4t zen
3. I could only find a comp between 845 and i3 from same reviewer.

 

[gamerk316]

People, you're going in circles here.

1: On IPC: As long as it reaches Sandy levels [and it almost certainly will], Ryzen will be *fast enough* to not be a bottleneck in gaming, which is all AMD really needs to do here. Anything extra is just gravy.

2: On SMT: We really can't guess how big the SMT gain is until we get benchmarks. Even then, we need a bench that tests both single and multi core, so we can do the math to isolate the SMT cores [something like: Single Core IPC * Number of physical cores, then using the remainder to calculate the SMT gains]. We'll know eventually. On paper, it should give a larger then the ~20% benefit HTT gives. I wouldn't be socked if the gains approach 50-60% honestly.

3: On overall performance: It's almost certain Intel will have higher IPC, and Ryzen will almost certainly win overall in multithreaded. The real question is, how close an i7 is to the corresponding Ryzen chip in multithreaded. Does the IPC benefit for Intel keep it close, or are we talking a 50% blowout? That answer will be the primary one, as it determines if anyone on an Intel platform has a real reason to switch over.

4: On IPC math in general: IPC is not a flat value; it fluctuates by workload. Don't try and compute a flat number; compare it across benchmarks and get an average. Could be IPC is competitive in Integer, less so in FP. How about SSE4/AVX/AVX2 workloads? And so on. There will be outliers, but you can still determine trends.

5: Chillaxe. We'll get benchies soon enough. You know I'm eagerly awaiting performing the math [I have a spreadsheet all ready to go; plug in the results and out comes IPC].

 

[sarwar_r87]

People, you're going in circles here.

1: On IPC: As long as it reaches Sandy levels [and it almost certainly will], Ryzen will be *fast enough* to not be a bottleneck in gaming, which is all AMD really needs to do here. Anything extra is just gravy.

2: On SMT: We really can't guess how big the SMT gain is until we get benchmarks. Even then, we need a bench that tests both single and multi core, so we can do the math to isolate the SMT cores [something like: Single Core IPC * Number of physical cores, then using the remainder to calculate the SMT gains]. We'll know eventually. On paper, it should give a larger then the ~20% benefit HTT gives. I wouldn't be socked if the gains approach 50-60% honestly.

3: On overall performance: It's almost certain Intel will have higher IPC, and Ryzen will almost certainly win overall in multithreaded. The real question is, how close an i7 is to the corresponding Ryzen chip in multithreaded. Does the IPC benefit for Intel keep it close, or are we talking a 50% blowout? That answer will be the primary one, as it determines if anyone on an Intel platform has a real reason to switch over.

4: On IPC math in general: IPC is not a flat value; it fluctuates by workload. Don't try and compute a flat number; compare it across benchmarks and get an average. Could be IPC is competitive in Integer, less so in FP. How about SSE4/AVX/AVX2 workloads? And so on. There will be outliers, but you can still determine trends.

5: Chillaxe. We'll get benchies soon enough. You know I'm eagerly awaiting performing the math [I have a spreadsheet all ready to go; plug in the results and out comes IPC].

we are going in circles as there are no more leaks, and we donot want this thread to die.

1. it will be ofay for us, not for average joe who makes up bulk of the market due to market perception.
2. even CMT does not give 60%. and that has a lot more transistor to back it

On average, Hyperthreading gives a gain of 23.4%, moving from four to six cores on K10 a gain of 42.2% and CMT in Bulldozer 53.1%. We’re therefore well beyond what you get with Hyperthreading and talking about Bulldozer as an 8-core architecture therefore does seem most accurate.

120% of 23% is still 27.3% gain in SMT for AMD.
4. no one said it is. what we are saying if its 50% better than previous gen in blender putting it in par with Haswell, we are saying it will be 50% better in other benchmark on average. BUT we are not saying zen will be Haswell level in other benchmarks.

EDIT> blender and handbrake are FP and Integer, so we know they are both more than compitative. question mark lies in AVX2

 

[sarinaide]

People, you're going in circles here.

1: On IPC: As long as it reaches Sandy levels [and it almost certainly will], Ryzen will be *fast enough* to not be a bottleneck in gaming, which is all AMD really needs to do here. Anything extra is just gravy.

2: On SMT: We really can't guess how big the SMT gain is until we get benchmarks. Even then, we need a bench that tests both single and multi core, so we can do the math to isolate the SMT cores [something like: Single Core IPC * Number of physical cores, then using the remainder to calculate the SMT gains]. We'll know eventually. On paper, it should give a larger then the ~20% benefit HTT gives. I wouldn't be socked if the gains approach 50-60% honestly.

3: On overall performance: It's almost certain Intel will have higher IPC, and Ryzen will almost certainly win overall in multithreaded. The real question is, how close an i7 is to the corresponding Ryzen chip in multithreaded. Does the IPC benefit for Intel keep it close, or are we talking a 50% blowout? That answer will be the primary one, as it determines if anyone on an Intel platform has a real reason to switch over.

4: On IPC math in general: IPC is not a flat value; it fluctuates by workload. Don't try and compute a flat number; compare it across benchmarks and get an average. Could be IPC is competitive in Integer, less so in FP. How about SSE4/AVX/AVX2 workloads? And so on. There will be outliers, but you can still determine trends.

5: Chillaxe. We'll get benchies soon enough. You know I'm eagerly awaiting performing the math [I have a spreadsheet all ready to go; plug in the results and out comes IPC].

To address IPC that is true, IPC is different depending on the application and how it uses resources. IPC is architectural and has to exist at hardware level to produce that amount of throughput. Per definition of factors affecting IPC

1) Clock speed
2) Memory controller.
3) Cache performance, Bulldozers CPU-NB resulted in cache being slower than DDR3, a complete and utter disaster.
4) Compute Units ie: Arithmetic units, Floating point units, queing registers etc etc Netburst and to lesser extent but still innefficient Bulldozer uArches were bad designs.

2) I have recently tested a Pentium G4560 with HTT and can tell you that it flat out beats a i5 2500K without stressing and anything less than Haswell bottlenecks a GTX1080 especially at higher resolutions. IPC performance can be linked to the BF demo it pushes very similar to the 6900K which means the CPU is capable at 3.4Ghz of pushing a Titan XP to the same frame rates without bottlenecking. A FX8370 cannot even make 20FPS at 4K while a 2600K is barely better still under acceptable 4K gaming with the CPU being the definite bottleneck.

If AMD just wanted sandy performance it would be complete lack of ambition and the fact that 89% of the market is Intel's nobody would even think of turning to AMD unless they flogged a 8C/16T ryzen for $300 which if it offered sandy performance would likely be the result in a couple months.

 

[Ziga Stupar]

well we do not know true IPC gains AMD was aiming at 40% increase over Excavator, but they claim they went past it so we are talking 40%+ IPC increase

 

[sarinaide]

Going back to CPC hardwares benches between productivity/synthetics and gaming I compared a i7 5960x with the 8370 due to similar clocks and then tried to match the games performance and it is very interesting.

CPC claim productivity is up about 60-65% while the gaming is up 35%

http://www.anandtech.com/bench/product/1709?vs=1317

That is a link to synthetics and some gaming.

Summary:

The aggregate difference in productivity is around 60-70%
The aggregate in gaming is between 30-40% depending if you take away BF4 which is heavily optimized for threading.

Games tested at 1080P with a GTX980ti Strix

Grid - 25%
Total War - 30%
Arma III - 53%
GTA V - 51%
Alien Isolation - 15%
Battlefield 4 - 8%

It was convenient that the i7 5960X was not part of that bench despite them mentioning similar performance.

My estimation is that AMD have produced higher IPC at lower clock speed and that has resulted in overall Haswell level performance. It is neither faster than the 6900K in productivity or gaming even with expected clock bumps and it was confirmed in the gaming demo that the 6900K was slightly faster in a well threaded game squeezing just a little bit more 5~ FPS out of the Titan XP.

 

[jdwii]

Spoiler

also it will have to perform too. because right now, you can get a 8 core for the same price as an i3. but you will hardly see anyone recommending a FX 8 core over i3. an IPC of SB will foot the bill, but will not help much as they will still be half a decade behind intel.

I do, sort of. Went from an FX 8120 to an i3-6100, and while the i3 is a lot faster in most stuff, it can't handle multitasking, to the point where there's stuttering with the mouse and sound on slightly heavy tasks. Anyone asking me today, I'd say: if you can't afford an i5, get an FX-6xxx. The system will be a lot smoother, games will run good, and the general experience will be better.

However, you are right in saying the issue is perception. Consumers prefer an i3 over any AMD. The job now is to change this perception, by pleasing enthusiasts first (they recommend hardware to people around them), by clever marketing (people must see AMD as a brand they can trust), and by good partnerships (like notebook builders).

I believe they can do it with IPC around Ivy Bridge-Haswell, because that's good enough, and they can play the value game again, but with a lot more margin this time. And if enthusiasts perceive value, they will sell value to their friends (who don't care about top performance, but value).

I actually don’t multitask much probably over me having a weak PC when I was younger that couldn’t even listen to MP3’s and browse the internet without locking up. But when it comes to gaming my 8350 was WAY worse than my 4360 back then my 770 actually had GPU usage around 70% most of the time with a 8350 when I jumped to a 4360 it was almost always at 95+%. Games I tested was Far cry 4, BF4, City Skylines, even farcry 3. I monitor my GPU usage and even with my new 1080 and 4790K I get 80% at times in GTA5 and some other games. Note that I own a 144hz G-sync monitor so it’s not that my GPU shouldn’t be at 100% usage until I get to that frame rate.

As for the little I do besides gaming I basically just watch videos and browse the web and my 8350 feels just like the I3 did and my I7.
I can’t help but wonder how badly a fx 6 core would bottleneck and I actually did test games using only 3 of the modules on my 8350 and games like BF4 played at even a lower frame rate and that was with a 770 only. IPC of piledriver-bulldozer is just to low to not bottleneck any modern mid-range GPU. Ryzen really couldn't come any sooner.

 

[sarinaide]

ZEN ES 3.1/3.3
i7 5960X 3/3.5 (3.3 all core)
FX 8370/50 4-4.4

Per Canard PC the ES is ~35% faster than the FX83XX

i7 5960X vs FX 83XX

All tests conducted with a GTX980ti at ultra settings on 1080P

Farcry 4 - 90/68 27%
Witcher 3 - 99/61 47%
Fallout 4 - 136/106 24%
GTA V - 97/57 51%
Shadow of Mordor - 160/103 43%
Arma III - 34/22 47%
Grid autosport - 175/137 25%

Test suite average = 37%

Productivity bench suite i7 5960X ~ 65-70% faster (Agisoft, Handbrake, H264, Dolphin, 7Zip per anandtech), the CPC benches showed Ryzen over a number of intel dominated test benches like wPrime and Corona having a ~60% advantage at lower clocks to the FX8370 which base+turbo is 25% higher than Ryzen. The only logical deduction is that AMD are trying not to show anything on this and later drop the details but there are to many pointers that performance is very good and that the Ryzen ES can drive the Titan XP.

Not quite 4K with only 25% but the sandybridges are clearly holding back the GTX1080 quite considerably compared to Haswell. I just can't see how AMD can replicate the results without significant IPC gains.

It is also impossible to just limit IPC gains to 40%, a minor change could yield that on it's own and IPC is extremely sensitive to changes. The economics of "just about sandybridge" performance is cataclysmic to AMD, it would be a bigger disappointment than Bulldozer with aging performance. Lisa Su is not the type of person that settles for inadequacy, doubt she would want to stake her reputation on barely passable sandy level performance in 2017

 

[Ziga Stupar]

ZEN ES 3.1/3.3
i7 5960X 3/3.5 (3.3 all core)
FX 8370/50 4-4.4

Per Canard PC the ES is ~35% faster than the FX83XX

i7 5960X vs FX 83XX

All tests conducted with a GTX980ti at ultra settings on 1080P

Farcry 4 - 90/68 27%
Witcher 3 - 99/61 47%
Fallout 4 - 136/106 24%
GTA V - 97/57 51%
Shadow of Mordor - 160/103 43%
Arma III - 34/22 47%
Grid autosport - 175/137 25%

Test suite average = 37%

Productivity bench suite i7 5960X ~ 65-70% faster (Agisoft, Handbrake, H264, Dolphin, 7Zip per anandtech), the CPC benches showed Ryzen over a number of intel dominated test benches like wPrime and Corona having a ~60% advantage at lower clocks to the FX8370 which base+turbo is 25% higher than Ryzen. The only logical deduction is that AMD are trying not to show anything on this and later drop the details but there are to many pointers that performance is very good and that the Ryzen ES can drive the Titan XP.

Not quite 4K with only 25% but the sandybridges are clearly holding back the GTX1080 quite considerably compared to Haswell. I just can't see how AMD can replicate the results without significant IPC gains.

It is also impossible to just limit IPC gains to 40%, a minor change could yield that on it's own and IPC is extremely sensitive to changes. The economics of "just about sandybridge" performance is cataclysmic to AMD, it would be a bigger disappointment than Bulldozer with aging performance. Lisa Su is not the type of person that settles for inadequacy, doubt she would want to stake her reputation on barely passable sandy level performance in 2017

so if i go by your benchmark chart and 35% increase over FX-83XX in Battlefield 1 it shows that Ryzen ES sample is 1-4 FPS below i5-6600, but current version has currently 500mhz higger base clock than tested ES not to mention boost so it suppased i5-6600 meybe even i5-6700

 

[sarinaide]

ZEN ES 3.1/3.3
i7 5960X 3/3.5 (3.3 all core)
FX 8370/50 4-4.4

Per Canard PC the ES is ~35% faster than the FX83XX

i7 5960X vs FX 83XX

All tests conducted with a GTX980ti at ultra settings on 1080P

Farcry 4 - 90/68 27%
Witcher 3 - 99/61 47%
Fallout 4 - 136/106 24%
GTA V - 97/57 51%
Shadow of Mordor - 160/103 43%
Arma III - 34/22 47%
Grid autosport - 175/137 25%

Test suite average = 37%

Productivity bench suite i7 5960X ~ 65-70% faster (Agisoft, Handbrake, H264, Dolphin, 7Zip per anandtech), the CPC benches showed Ryzen over a number of intel dominated test benches like wPrime and Corona having a ~60% advantage at lower clocks to the FX8370 which base+turbo is 25% higher than Ryzen. The only logical deduction is that AMD are trying not to show anything on this and later drop the details but there are to many pointers that performance is very good and that the Ryzen ES can drive the Titan XP.

Not quite 4K with only 25% but the sandybridges are clearly holding back the GTX1080 quite considerably compared to Haswell. I just can't see how AMD can replicate the results without significant IPC gains.

It is also impossible to just limit IPC gains to 40%, a minor change could yield that on it's own and IPC is extremely sensitive to changes. The economics of "just about sandybridge" performance is cataclysmic to AMD, it would be a bigger disappointment than Bulldozer with aging performance. Lisa Su is not the type of person that settles for inadequacy, doubt she would want to stake her reputation on barely passable sandy level performance in 2017

so if i go by your benchmark chart and 35% increase over FX-83XX in Battlefield 1 it shows that Ryzen ES sample is 1-4 FPS below i5-6600, but current version has currently 500mhz higger base clock than tested ES not to mention boost so it suppased i5-6600 meybe even i5-6700

Dresdenboy recalculated the clockspeed to 3.4/3.8 and Ryzen would score between the 6800 and 6900K at around 106% vs 98% where it is rated now, that is about 11% base clock and 14% turbo uplift over the ES used by them. The AMD test bench had a 3.4Ghz locked Ryzen ES vs i7 6900K 3.3 base, 3.5 all core and 3.7 1 core turbo running 1-2% slower FPS which is about 2-5FPS

Just for fun comparing gaming on a i5 6600 to a 5960 shows that the 5960X is slower in less threaded games where Skylake architecture is better than Broadwell and the slightly higher clocks help.

 

[jdwii]

Since that is BF1 none of those tests measure IPC if however you used Dolphin benchmark or something like that i'd say otherwise.

Soon we will see the true IPC numbers soon. Even Cities skylines would have been a better example.

Edit
http://www.anandtech.com/show/8316/amds-5-ghz-turbo-cpu-in-retail-the-fx9590-and-asrock-990fx-extreme9-review/6