Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)

[DisEnchantment]

Speculate at will

 

[SarahKerrigan]

AMD hasn't lied or missed a single DC part in Zen history so this would be a first.
Everything since Rome was exactly as promised.

And yet your proclaimed +32% has been proven to be wrong, and the MLID slide's "10-15%+" has been proven to be right.

What are you actually arguing here?

 

[poke01]

remember 2022 computex Zen 4, none of this AI BS

 

[adroc_thurston]

Dude, you got duped. Get over it.

Tier 1 OEMs do not dupe people.

And yet your proclaimed +32% has been proven to be wrong, and the MLID slide's "10-15%+" has been proven to be right.

What are you actually arguing here?

Trying to reconcile the data.

 

[SarahKerrigan]

Tier 1 OEMs do not dupe people.

You should have seen what customers were hearing from HP/Intel before Tukwila and Poulson releases.

 

[gdansk]

remember 2022 computex Zen 4, none of this AI BS

There was less but 2022 wasn't that great

 

[adroc_thurston]

You should have seen what customers were hearing from HP/Intel before Tukwila and Poulson releases.

Yeah I know the old days, but neither Xeon nor EPYC side was wrong in projections in many years.

 

[Saylick]

Dang, yet another partner on the stage...

For the record, I'm still drinking my beer, but it's not even because of AI, it's because I'm getting bored. So bored that I'm doing work on the side right now... got some stuff I needed to catch up for this upcoming week.

 

[GTracing]

Impressive numbers, it will be interesting to see how the X Eilte and Strix compare in independent testing.

 

[Jayzen]

Maybe the "40%" figure was about NPU performance this entire time. Maybe that's how Kepler/Adroc will save themselves.

 

[StinkyPinky]

The numbers mean little if we don't fully know the power draw.

 

[Tuna-Fish]

These partner showcases are going to be the death of me. AI per sentence.

 

[IEC]

Should be thankful Microsoft didn't push for NPU's on the desktop.

That "feature" needs to be totally... recalled

I think 16% is underwhelming, and deff worse than Zen 4, but I also don't think it's atrocious.
Also, do you guys notice that the perf uplift in gaming seems to be... weird? Their comparison vs the 14900k (at least from that videocardz leak) is like 13% better on average, which would mean that it's 25-30% better than Zen 4.

Frequency while gaming will matter. My 7700X may have a 5.4GHz boost clock but it was lucky to be hitting ~5GHz in gaming workloads. If they combine +16% with at least some frequency gain that could explain the difference.

 

[adroc_thurston]

Maybe the "40%" figure was about NPU performance this entire time. Maybe that's how Kepler/Adroc will save themselves.

No.
I'm talking strictly vendor-communicated server perf in SIR scores.

 

[gdansk]

The numbers mean little if we don't fully know the power draw.

Isn't this always the case for every new laptop chip

 

[poke01]

Dang, yet another partner on the stage...

For the record, I'm still drinking my beer, but it's not even because of AI, it's because I'm getting bored. So bored that I'm doing work on the side right now... got some stuff I needed to catch up for this upcoming week.

so many

View attachment 100277
Impressive numbers, it will be interesting to see how the X Eilte and Strix compare in independent testing.

did AMD really compare HX370 to the base M3? Not even the M3 Pro. Apple might be the laptop leader this year with the M4 Pro/Max

 

[Abwx]

Perf improvement is still largely enough in respect of the present and future competition, on the mobile department that s a no contest, a Core Ultra 185H has to be crancked at roughly 90W to barely match a 12 cores Strix Point at 35W, no wonder that they got that much design wins.

 

[Makaveli]

Lol Asus guy seems so impressed with himself.

 

[Saylick]

That "feature" needs to be totally... recalled

Lol.

Frequency while gaming will matter. My 7700X may have a 5.4GHz boost clock but it was lucky to be hitting ~5GHz in gaming workloads. If they combine +16% with at least some frequency gain that could explain the difference.

I doubt it. Wider cores don't normally miraculously clock higher with a relatively unchanged node.

 

[Ghostsonplanets]

View attachment 100277
Impressive numbers, it will be interesting to see how the X Eilte and Strix compare in independent testing.

Do we have 3DMark Night Raid data for Snap X Elite? I want to know what 1.6x improvement means over it.

 

[Tuna-Fish]

And I'm done. ASUS did it.

 

[Saylick]

Jeez... Since when did AMD stand for ASUS Mobile Devices?

 

[GTracing]

Do we have 3DMark Night Raid data for Snap X Elite? I want to know what 1.6x improvement means over it.

I don't put much stock in that comparison. Synthetic graphics performance doesn't correlate well with real gaming performance.

 

[StinkyPinky]

Isn't this always the case for every new laptop chip

Yeah, but I note they don't include that in any of their slides so it guzzles more juice than the Elite which probably means more fan noise and less battery time.

 

[poke01]

Do we have 3DMark Night Raid data for Snap X Elite? I want to know what 1.6x improvement means over it.

no need, X Elite has a worse GPU than a Rog ally

 

[RnR_au]

Strix Point laptops available in July... noice!