AMD 8-core ZEN Packs a Whallop with Multithreaded Performance

[sergionography]

Wouldnt this be saying the OPPOSITE?

In order for a 8 core zen to outperform a 10 core xeon at multithreaded benchmarks, assuming the core clock is similar, the zen cores would need to be more powerful then the i cores at the same speed, which would indicate better single threaded performance.

Intel 10 core at 2.6ghz
Amd 8 core at ???

My assumption amd is clocked at 3.2ghz+

10 × 2.6 = 26
26 ÷ 8 = 3.25ghz

Sounds in line with expectations. amd doesnt need to beat intel in single thread they just need to match them or be close enough within 15%ipc to compete because the rest they can maneuver between clock speed amd core count.


Also for those wondering why all the benchmarks and hype is around multithreaded its because thats logically where all the ironing out part comes as far as engineering samples go. because its easy to clock one core to max but its definitely more tricky to iron out the thread management and power delivery and fine tuning for the chip as a whole, and thats what decides the total performance of the overall package and what will allow single cores to clock higher when running single thread

 

[Camm]

I would also add that we are now moving into a post single core era. Take BF1's scaling across multiple cores for example.

My projection is that AMD will have about a 10% IPC deficit against Intel, but will counter that with aggressive core count pricing. Whilst postulating about server pricing might be a bit rich, the easiest example I can think of is pushing an 8 core part against 1151 4 core parts.

Still, hoping for some competition!

 

[Scrizz]

The Xeon E5-2650 V2 is incorrectly quoted in this article as having 10 cores, when it actually has 8 (http://ark.intel.com/products/75269/Intel-Xeon-Processor-E5-2650-v2-20M-Cache-2_60-GHz).
Perhaps the author was confusing it with the V3 variant, or the E5-2660 V2, both of which do have 10 cores?

The Xeon E5-2680 V2 (which performs better than the Zen ES in this benchmark) has 10 cores, so Zen seems to be between an 8 and 10 core Ivy Bridge-EP Xeon, which in multi-thread aware benchmarks (like this one) is where a 6 core i7-6850K would place. This is probably not the conclusion everyone else in this thread was getting by going off that initial mistake in the article.

Edit: This benchmark seems a bit off. The must be some other variable at play, otherwise there is no way a Haswell-EP 2697-V3 (14 cores) could lose to an Ivy Bridge-EP 2690-V2 (10 cores) so badly.

Thank you!

 

[Daniel L]

Let's hope this is really, really, false. The panic alarm here seems to be the devastating chart. Why are we comparing this AMD engineering sample with Xeons from 2013? This AMD engineering sample is placed in the cart right between a Xeon E5-2680 v2 2.8GHz and Xeon E5-2650 v2 2.6GHz; both more than 3 years old. I want parts that compete with Intel parts now, not Intel parts of 2013.

Intel Xeon E5-2658 v2 @ 2.40GHz
PassMark scores 13,875

Intel Core i7-6850K @ 3.60GHz
PassMark scores 14,318

Intel Core i7-5960X @ 3.00GHz
PassMark scores 15,974

[URL='http://www.cpubenchmark.net/cpu.php?cpu=Intel+Xeon+E5-2680+v2+%40+2.80GHz&id=2061']Intel Xeon E5-2680 v2 @ 2.80GHz
PassMark scores 16,341

Intel Core i7-6900K @ 3.20GHz
PassMark scores 17,431

Intel Core i7-6950X @ 3.00GHz
PassMark scores 20,018[/URL]
passmark score wise that would put the AMD zen engineering sample (+\-) Intel Core i7-6850K @ 3.60GHz as a true competitor

 

[Daniel L]

Thank you!

probabbly
Intel Xeon E5-2658 v2 @ 2.40GHz 10 core

 

[Captain_Tom]

I would also add that we are now moving into a post single core era. Take BF1's scaling across multiple cores for example.

My projection is that AMD will have about a 10% IPC deficit against Intel, but will counter that with aggressive core count pricing. Whilst postulating about server pricing might be a bit rich, the easiest example I can think of is pushing an 8 core part against 1151 4 core parts.

Still, hoping for some competition!

The thing is AMD doesn't need to win in ultimate performance. If a 28-core Xeon edges out the 32-core Naples - it really doesn't matter.


What will matter is if AMD can crush Intel in Price/perf while simultaneously having better efficiency. The leak showing a 95w Zen CPU beating a 140w i7-6900K showed that this could be a reality. It would be a long overdue upset in the CPU space.

 

[bug]

The thing is AMD doesn't need to win in ultimate performance. If a 28-core Xeon edges out the 32-core Naples - it really doesn't matter.


What will matter is if AMD can crush Intel in Price/perf while simultaneously having better efficiency.

Actually, you couldn't be more wrong. After all these years in the red, what AMD needs is a cash cow. For that, they actually need to beat intel, not get bogged down in price wars.

 

[TheLaughingMan]

Momma and daddy stop fighting.

 

[BlakSlabeth]

@Assimilator

Maybe you're too young to remember when both Intel and AMD produced competitive products - REAL competition. Perform gains between generations were significant. I *need* more than 4 cores. Here in Canada, I'm looking at CPU costs for an Intel 6 core 12 thread - 6850 - $840. 8 core - almost $1400. That's just the CPU. Can't wait for AMD to bring competition back. I'll happily pay $400 for 8c chip when each core is 80+% of the intel equivalent.

 

[Captain_Tom]

Actually, you couldn't be more wrong. After all these years in the red, what AMD needs is a cash cow. For that, they actually need to beat intel, not get bogged down in price wars.

I mean what I said would effectively be "beating" Intel. There are 3 ways to "win": Price/perf, Ultimate Perf, and Efficiency. If AMD wins 2 out of those 3 they will sell a ton of hardware and rake in the money.

We don't know if they will yet, but it looks like they might.

 

[m1dg3t]

@Assimilator

Maybe you're too young to remember when both Intel and AMD produced competitive products - REAL competition.

So real that Intel paid OEMs to not use AMD products. The good ole days...

 

[TheGuruStud]

So real that Intel paid OEMs to not use AMD products. The good ole days...

I thought making their socket 7 MBs crash with AMD chips was the good ol days?

 

[Captain_Tom]

So real that Intel paid OEMs to not use AMD products. The good ole days...

I don't think AMD will hit as big as a homerun as they did with the original FX series, but at the same time Intel can't pull the same shady tactics anymore (I hope).


AMD may actually be able to do better with a relatively worse product if the the laws and public opinion are on their side.

 

[ensabrenoir]

.......seems familiar somehow

....ah yes celebrating too early.......

 

[bug]

I mean what I said would effectively be "beating" Intel. There are 3 ways to "win": Price/perf, Ultimate Perf, and Efficiency. If AMD wins 2 out of those 3 they will sell a ton of hardware and rake in the money.

We don't know if they will yet, but it looks like they might.

Actually, the three ways to win are price, performance and power consumption (price/performance is just a combination of the first two). For their own good, they'd better win in the last two categories.

 

[Assimilator]

@Assimilator

Maybe you're too young to remember when both Intel and AMD produced competitive products - REAL competition. Perform gains between generations were significant. I *need* more than 4 cores. Here in Canada, I'm looking at CPU costs for an Intel 6 core 12 thread - 6850 - $840. 8 core - almost $1400. That's just the CPU. Can't wait for AMD to bring competition back. I'll happily pay $400 for 8c chip when each core is 80+% of the intel equivalent.

Nope, I'm plenty old enough to remember the glory days of the P4 getting slaughtered left, right and center by the Athlon 64. The first PC I built when I got a job was an Athlon 64 system... I still have the Asus A8N32-SLI Deluxe board from that one!

People seem to assume that I hate AMD, or want them to fail. But that's the opposite of the truth. I would love if AMD could bring us back the Athlon 64 days, when you didn't have to pay extra for the ability to overclock your CPU and a new generation meant more than single-digit IPC increases.

But the AMD of now isn't the AMD that delivered the Athlon 64. That company was run by engineers, this company is run by suits whose solution to engineering problems is to spend more money on marketing. I absolutely despise deceptive practices, and AMD's marketing since Bulldozer has been lie after lie; until they actually deliver the product they're selling, I refuse to trust any "leaks" that claim their products will beat Intel. Sue me for being a cynic.

I want the old AMD back - the AMD that put their money where their mouth was and didn't deliver products that were merely good enough. That AMD would've looked at these benchmarks as a failure, because competing with years-old products isn't competing at all. That AMD would have done better.

 

[TheGuruStud]

I don't think AMD will hit as big as a homerun as they did with the original FX series, but at the same time Intel can't pull the same shady tactics anymore (I hope).


AMD may actually be able to do better with a relatively worse product if the the laws and public opinion are on their side.

I want the old AMD back - the AMD that put their money where their mouth was and didn't deliver products that were merely good enough. That AMD would've looked at these benchmarks as a failure, because competing with years-old products isn't competing at all. That AMD would have done better.

Losing untold billions due to intel being afraid to legally compete has finished that possibility off. They got off with a slap on the wrist. Intel will do it, again, if necessary. There's no incentive to stop (just like ignoring safety, enviromental, etc regulations).

 

[Captain_Tom]

Nope, I'm plenty old enough to remember the glory days of the P4 getting slaughtered left, right and center by the Athlon 64. The first PC I built when I got a job was an Athlon 64 system... I still have the Asus A8N32-SLI Deluxe board from that one!

People seem to assume that I hate AMD, or want them to fail. But that's the opposite of the truth. I would love if AMD could bring us back the Athlon 64 days, when you didn't have to pay extra for the ability to overclock your CPU and a new generation meant more than single-digit IPC increases.

But the AMD of now isn't the AMD that delivered the Athlon 64. That company was run by engineers, this company is run by suits whose solution to engineering problems is to spend more money on marketing. I absolutely despise deceptive practices, and AMD's marketing since Bulldozer has been lie after lie; until they actually deliver the product they're selling, I refuse to trust any "leaks" that claim their products will beat Intel. Sue me for being a cynic.

I want the old AMD back - the AMD that put their money where their mouth was and didn't deliver products that were merely good enough. That AMD would've looked at these benchmarks as a failure, because competing with years-old products isn't competing at all. That AMD would have done better.

I mean when Intel (And more recently Nvidia) resort to shady or dishonest tactics AMD kinda is forced to follow suit.

Don't get me wrong, I don't think AMD was ever a saint, nor do I think they are that dishonest lately. But I have to admit the leaks around Bulldozer were complete BS, but can you blame them? They had to bend the truth to move product against a wholly dishonest competitor.

 

[bug]

Nope, I'm plenty old enough to remember the glory days of the P4 getting slaughtered left, right and center by the Athlon 64. The first PC I built when I got a job was an Athlon 64 system... I still have the Asus A8N32-SLI Deluxe board from that one!

People seem to assume that I hate AMD, or want them to fail. But that's the opposite of the truth. I would love if AMD could bring us back the Athlon 64 days, when you didn't have to pay extra for the ability to overclock your CPU and a new generation meant more than single-digit IPC increases.

But the AMD of now isn't the AMD that delivered the Athlon 64. That company was run by engineers, this company is run by suits whose solution to engineering problems is to spend more money on marketing. I absolutely despise deceptive practices, and AMD's marketing since Bulldozer has been lie after lie; until they actually deliver the product they're selling, I refuse to trust any "leaks" that claim their products will beat Intel. Sue me for being a cynic.

I want the old AMD back - the AMD that put their money where their mouth was and didn't deliver products that were merely good enough. That AMD would've looked at these benchmarks as a failure, because competing with years-old products isn't competing at all. That AMD would have done better.

I remember those days as well. Though AthlonXP was the first to wipe the floor with P4, but you had to overclock the Athlon for that. K6-2 and K6-3 were no slouches either, they were cheaper alternatives to PII and P3, as long as you didn't need floating point performance. Also, the AthlonXP and Athlon64 aren't 100% AMD's success, they were the result of AMD poaching talent (and probably IP - even Wikipedia is oblivious to this) from DEC.
And while that AMD was run by engineers, it was criticized back then for basically not having any significant marketing to speak of. It's also that AMD that was totally unprepared for intel's shady tactics, even though a kid could have seen that coming. Yes, the engineer in me is rooting for that AMD, but realistically speaking they need both engineering and marketing prowess. Unfortunately they're out of cash, so I wouldn't want to be in the shoes of anyone running AMD these days.

 

[cdawall]

I remember those days as well. Though AthlonXP was the first to wipe the floor with P4, but you had to overclock the Athlon for that. K6-2 and K6-3 were no slouches either, they were cheaper alternatives to PII and P3, as long as you didn't need floating point performance. Also, the AthlonXP and Athlon64 aren't 100% AMD's success, they were the result of AMD poaching talent (and probably IP - even Wikipedia is oblivious to this) from DEC.
And while that AMD was run by engineers, it was criticized back then for basically not having any significant marketing to speak of. It's also that AMD that was totally unprepared for intel's shady tactics, even though a kid could have seen that coming. Yes, the engineer in me is rooting for that AMD, but realistically speaking they need both engineering and marketing prowess. Unfortunately they're out of cash, so I wouldn't want to be in the shoes of anyone running AMD these days.

Athlon XP didn't mop the floor with P4. Look at the timeframes again and then look at some benchmarks.

 

[bug]

Athlon XP didn't mop the floor with P4. Look at the timeframes again and then look at some benchmarks.

Ok, bad choice of words. It could match a P4 for about half the money. IIRC, initially AthlonXP went against Northwood and then Athlon64 went against Prescott (aka Preshot).

Edit: Come to think of it, everybody's had a lemon at some point. Intel had P4, Nvidia had their FX5000 series. And AMD... AMD got themselves into a position worse than David's against Goliath.

 

[xorbe]

Intel was aware that P4 was clocking too fast and too hot. As far as I know, Prescott came out because the replacement wasn't ready yet. AMD took far too long to accept that the Bulldozer line wasn't cutting it as a mainstream part. Hopefully Zen at least rights their listing ship, even if it doesn't achieve speedboat results.

 

[TheGuruStud]

Athlon XP didn't mop the floor with P4. Look at the timeframes again and then look at some benchmarks.

I also recall most review publications being incompetent or intel shills and pushing tons of synthetic benchmarks (damage control). Remember the case of one owned by intel (sysmark?) producing results of intel winning hands down? Those lies even made it into textbooks. I'm sure intel's compiler took care of the rendering scores.
And at more than twice the cost, intel was a huge loser, but they sure struck back with their little penis killing AMD's market share under the best x86 CPU ever made.

 

[cdawall]

I also recall most review publications being incompetent or intel shills and pushing tons of synthetic benchmarks (damage control). Remember the case of one owned by intel (sysmark?) producing results of intel winning hands down? Those lies even made it into textbooks. I'm sure intel's compiler took care of the rendering scores.
And at more than twice the cost, intel was a huge loser, but they sure struck back with their little penis killing AMD's market share under the best x86 CPU ever made.

Even gaming benchmarks showed the 2.4's competing comfortably with AMD's top end at the time. The athlon XP 3200+ competed with the 3.2C, but it didn't wipe the floor with it in games and synthetics they were just competitive it wasn't until the Athlon 64 came out that they literally wiped the floor with everything like it was some kind of cruel joke.

 

[nemesis.ie]

But the AMD of now isn't the AMD that delivered the Athlon 64. That company was run by engineers, this company is run by suits whose solution to engineering problems is to spend more money on marketing. I absolutely despise deceptive practices, and AMD's marketing since Bulldozer has been lie after lie; until they actually deliver the product they're selling, I refuse to trust any "leaks" that claim their products will beat Intel. Sue me for being a cynic.

I want the old AMD back -

I think that situation is changing, Lisa Su is (AFAIK) an engineer. Also, I think part of AMD's issue when they were very competitive was actually a lack of marketing. So if they can restore the balance; get out good products but also make sure people know about them then things will improve.

I started a course recently and 90% of the folks in it who have been looking for "back to school" laptops are purely asking the question "Should I buy an I5 or I7" they have no clue about the rest of the system nor balance nor the importance of a good screen etc.

So the marketing dept definitely has a place; if it is used to (honestly) educate the consumers it's all good IMO.