The Impact Of GCC Zen Compiler Tuning On AMD Ryzen Performance

[zboson]

Interesting. So Zen will be more competitive with Intel on tasks that are less memory-intensive and more cache efficient? Which is why it seems to perform poorly on some games?

Zen is also dual channel if I recall whereas Skylake (not sure which was first) is quad-channel. This means Zen is more affected by memory bandwidth. That's maybe the second most disappointing thing about Zen after sticking with AVX128. I'm still likely to build a Zen system. It will be the first desktop I have build in years.

[Shimon]

< Stay tuned!

Nice pun!

[Holograph]

Zen is also dual channel if I recall whereas Skylake (not sure which was first) is quad-channel. This means Zen is more effected by memory bandwidth. That's maybe the second most disappointing thing about Zen after sticking with AVX128. I'm still likely to build a Zen system. It will be the first desktop I have build in years.

Intel only has quad-channel in the socket 2011 stuff. They have never done quad-channel in a mainstream consumer product. And Skylake for 2011 isn't out.

As for the article... hoping GCC7 takes this further. Fairly disappointing. But hey, I'm even more disappointed in the platform itself with its extreme lack of PCIe and SATA connectivity.

[SaucyJack]

Zen is also dual channel if I recall whereas Skylake (not sure which was first) is quad-channel. This means Zen is more affected by memory bandwidth. That's maybe the second most disappointing thing about Zen after sticking with AVX128. I'm still likely to build a Zen system. It will be the first desktop I have build in years.

With the exception of some very niche applications there is no performance difference between 2400mhz dual channel ddr4 and anything higher.

[zboson]

Intel only has quad-channel in the socket 2011 stuff. They have never done quad-channel in a mainstream consumer product. And Skylake for 2011 isn't out.

As for the article... hoping GCC7 takes this further. Fairly disappointing. But hey, I'm even more disappointed in the platform itself with its extreme lack of PCIe and SATA connectivity.

Good point. I read about the 6900K and generalized it too far. So once of the advantages the twice as expensive 6900K still has over Ryzen is quad channel.
https://arstechnica.com/gadgets/2017...t-can-compete/

"One particular Geekbench subtest showed a strong advantage in the other direction. Geekbench includes tests of the cryptographic instructions found in all mainstream processors these days. In a test of single-threaded performance, the Ryzen trounces the Broadwell-E, encrypting at 4.5GB/s compared to 2.7GB/s. Ryzen has two AES units that both reside within the floating point portion of the processor. Broadwell only has one, giving the AMD chip a big lead.

This situation is reversed when moving from one thread to 16: the Intel system can do 24.4GB/s while the AMD only does 10.2GB/s. This suggests that the test becomes bandwidth-limited with high thread counts, allowing the 6900K's quad-memory channels to give it a lead over the 1800X's dual channels. Even though the Ryzen has more computational resources to throw at this particular problem, that doesn't help when the processor sits twiddling its thumbs waiting for data."

[Holograph]

I'd be curious to know if overclocking the memory on Ryzen helps, but I know Michael isn't really into overclocking. And supposedly a lot of people are having issue trying to overclock memory with Ryzen so far, but some have pulled it off. I'm hoping more people will release data on this. Of course this question is worthless to anyone who is using Linux for productivity-related stuff.

[Michael]

I'd be curious to know if overclocking the memory on Ryzen helps, but I know Michael isn't really into overclocking. And supposedly a lot of people are having issue trying to overclock memory with Ryzen so far, but some have pulled it off. I'm hoping more people will release data on this. Of course this question is worthless to anyone who is using Linux for productivity-related stuff.

I've had no luck getting the memory speed ramped higher, will try when new firmware is out

[Holograph]

I've had no luck getting the memory speed ramped higher, will try when new firmware is out

Awesome. Thanks for trying.

[edwaleni]

"One particular Geekbench subtest showed a strong advantage in the other direction. Geekbench includes tests of the cryptographic instructions found in all mainstream processors these days. In a test of single-threaded performance, the Ryzen trounces the Broadwell-E, encrypting at 4.5GB/s compared to 2.7GB/s. Ryzen has two AES units that both reside within the floating point portion of the processor. Broadwell only has one, giving the AMD chip a big lead.

This situation is reversed when moving from one thread to 16: the Intel system can do 24.4GB/s while the AMD only does 10.2GB/s. This suggests that the test becomes bandwidth-limited with high thread counts, allowing the 6900K's quad-memory channels to give it a lead over the 1800X's dual channels. Even though the Ryzen has more computational resources to throw at this particular problem, that doesn't help when the processor sits twiddling its thumbs waiting for data."

Excellent analysis and could explain some behaviors. Thank you for sharing.

[anth]

According to an analysis of Ryzen cache by hardware.fr (which Chrome translates well): latency is very poor when a core from one of the two four-core-clusters accesses something in the L3 cache of the other.