Question Intel Mont thread

Page 17 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

DavidC1

Golden Member
Dec 29, 2023
1,452
2,361
96
Just to be sure I understood.
AVX is desirable, but in practice it's just a way to execute the same calculations that sse do, right? And because Tremont greatly improved SSE the performance uplift wouldn't be that different if it had AVX1, or something like that, right?
Gracemont is way better than Tremont, because of the much better architecture. AVX supports new instructions that SSE doesn't and in certain cases it will be better. The thing is though vast majority of applications are complex and won't benefit immensely even if it had support and took advantage of such instructions.

If Gracemont is roughly on Skylake level per clock then Tremont is Ivy Bridge.
 

DavidC1

Golden Member
Dec 29, 2023
1,452
2,361
96
And how about the consumer tier? Just go up to octa core?
May not even have a dedicated version at all. The successor seems to be "Wildcat Lake" which is essentially the smallest version of the Panther Lake mainstream platform:

That's 2P + 4 "LPE"(similar to Lunarlake).
 

DavidC1

Golden Member
Dec 29, 2023
1,452
2,361
96
It was created because AVX(2) were rudimentary ISAs.
And I repeat, it should have been new instructions with the same 256-bit width. It doesn't need 512-bit. I bet you if AVX-512 was instead AVX-256, Skymont would have supported it with double pumping and there would have been no fragmentation. Fragmentation exists because the Skymont team is actually focused on efficiency.
Just don't.
I'll NEVER agree with the people who believe needing recompile every 2 years to support a wider vector set is not just recommended, but necessity.

And I'm glad the E core team and the ARM guys agree on this. Increasing number of FP units benefits everything 30+ years.

Sometimes it seems the IT industry believes disrupting everything and anything is a necessity, but it really is an elitist mindset and benefits primarily themselves.
 
Reactions: igor_kavinski

adroc_thurston

Diamond Member
Jul 2, 2023
5,681
7,899
96
And I repeat, it should have been new instructions with the same 256-bit width.
no.
I'll NEVER agree with the people who believe needing recompile every 2 years to support a wider vector set is not just recommended, but necessity.
You really need to talk to some actual real SIMD people.
And I'm glad the E core team and the ARM guys agree on this
Well neither will ever have any SIMD code running on them.
Increasing number of FP units benefits everything 30+ years.
no it didn't, FMA-wise we went from 1 to 2 where it matters and that's it.
Even Knights* thingamajigs had 2 FMA units for very obvious reasons.
 

DZero

Golden Member
Jun 20, 2024
1,084
390
96
May not even have a dedicated version at all. The successor seems to be "Wildcat Lake" which is essentially the smallest version of the Panther Lake mainstream platform:

That's 2P + 4 "LPE"(similar to Lunarlake).
Damn.... so those 2P have HT or not?
 

DavidC1

Golden Member
Dec 29, 2023
1,452
2,361
96
Thanks for confirming that you are just huffing hopium
Wildcat Lake is 2P+4LPE of the next generation cores. The "LPE" cores are likely arranged in a similar way to Lunarlake, where it actually adds significant multi-threaded performance while still being able to scale to very low power, unlike Meteorlake's useless LPE which could barely power local video playback.

2P is 2x Cougar Cove, which won't be a huge gain over Lion Cove at probably 6-8% gains, but it is a huge improvement over Skylake-class cores.

Darkmont is a small(2-4%) improvement over the famed Skymont cores, which is drastically faster than the Gracemont cores.

So in the all-important single thread, Cougar Cove will annihilate 8E cores based on Gracemont, and despite lacking HT, 2P + 4LPE of greatly faster cores will be faster in MT as well.

Golden Cove = 40-50% faster per clock over Gracemont
Lion Cove = ~10% faster
Cougar Cove = ~8% faster
Plus higher clocks

You are talking 3.5+ Gracemont cores in the P setup alone. Then the 4LPE Darkmont cores should add at least 5.5 on top of that. We're already 10%+ faster than the predecessor.
 
Reactions: DAPUNISHER

eek2121

Diamond Member
Aug 2, 2005
3,339
4,914
136
Wildcat Lake is 2P+4LPE of the next generation cores. The "LPE" cores are likely arranged in a similar way to Lunarlake, where it actually adds significant multi-threaded performance while still being able to scale to very low power, unlike Meteorlake's useless LPE which could barely power local video playback.

2P is 2x Cougar Cove, which won't be a huge gain over Lion Cove at probably 6-8% gains, but it is a huge improvement over Skylake-class cores.

Darkmont is a small(2-4%) improvement over the famed Skymont cores, which is drastically faster than the Gracemont cores.

So in the all-important single thread, Cougar Cove will annihilate 8E cores based on Gracemont, and despite lacking HT, 2P + 4LPE of greatly faster cores will be faster in MT as well.

Golden Cove = 40-50% faster per clock over Gracemont
Lion Cove = ~10% faster
Cougar Cove = ~8% faster
Plus higher clocks

You are talking 3.5+ Gracemont cores in the P setup alone. Then the 4LPE Darkmont cores should add at least 5.5 on top of that. We're already 10%+ faster than the predecessor.
The issue with these parts are always clocks. I want wanted a small system with *mont cores only, but Intel clocks them really low compared to hybrid parts.

I guess they are worried about cannibalizing bigger chip sales.
 

DavidC1

Golden Member
Dec 29, 2023
1,452
2,361
96
The issue with these parts are always clocks. I want wanted a small system with *mont cores only, but Intel clocks them really low compared to hybrid parts.
The chips with mont cores only also happen to be very low TDP, which would limit the clocks as well.

Lunarlake's E cores clock only at 3.7GHz, which is not very high either.
 

511

Platinum Member
Jul 12, 2024
2,149
1,859
106
The issue with these parts are always clocks. I want wanted a small system with *mont cores only, but Intel clocks them really low compared to hybrid parts.

I guess they are worried about cannibalizing bigger chip sales.
Yeah same they would be able to easily fit 8 Darkmont E cores on ring with good ST vs Very Good Multi vs 2P+4LPE Very good ST but OK Multi.

Thanks David for explaining this also let's not forget 3 node shrinks as well.
 

LightningZ71

Platinum Member
Mar 10, 2017
2,195
2,676
136
Remember, an all darkmont processor need not have crippled ST throughput. With things like fin-flex, one cluster can be flexed for higher clocks and the rest can be flexed for efficiency. They could even choose to scribe the cluster with less density if they really wanted.
 

511

Platinum Member
Jul 12, 2024
2,149
1,859
106
Remember, an all darkmont processor need not have crippled ST throughput. With things like fin-flex, one cluster can be flexed for higher clocks and the rest can be flexed for efficiency. They could even choose to scribe the cluster with less density if they really wanted.
Darkmont is 18A there is no NanoFlex as for crippled ST Performance it will be around slightly higher than Raptor Cove IPCand judging by E and LP-E 4.5 GHz should be achievable.
 
Last edited:

LightningZ71

Platinum Member
Mar 10, 2017
2,195
2,676
136
Darkmont is 18A there is no NanoFlex as for crippled ST Performance it will be around slightly higher than Raptor Cove IPCand judging by E and LP-E 4.5 GHz should be achievable.
I must have misread that someplace; I stand corrected. Still, just as AMD did for their Version of a hybrid architecture, it's possible to do design and implementation stage circuit layout optimizations and even throw extra transistors into the mix to allow a given logical design to clock higher.
 
Reactions: Tlh97 and 511

511

Platinum Member
Jul 12, 2024
2,149
1,859
106
I must have misread that someplace; I stand corrected. Still, just as AMD did for their Version of a hybrid architecture, it's possible to do design and implementation stage circuit layout optimizations and even throw extra transistors into the mix to allow a given logical design to clock higher.
Yeah AMD uses clear tricks like changing the Libs and metal tracks while keeping the arch mostly the same though they sacrifice clock cycles for this.
 

DavidC1

Golden Member
Dec 29, 2023
1,452
2,361
96
Intel isn't gonna bother with such tricks on such a low end part. Also, if the TDP is low that's going to keep all core clocks relatively low as well. Still I do expect it'll be a faster CPU overall than Alderlake/Twinlake-N.
 
Reactions: 511
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |