Generic CPU discussion

Tom M
Tom M
Joined: 2 Feb 06
Posts: 6475
Credit: 9603371617
RAC: 5781203

https://www.anandtech.com/sho

https://www.anandtech.com/show/18764/fujitsu-preps-monaka-datacenter-cpu-to-succeed-a64fx-higher-performance-and-more-features

And has a link if you wanted to buy your very own A64fx. I think it was "on sale" from a vendor starting at $38,000

A Proud member of the O.F.A.  (Old Farts Association).  Be well, do good work, and keep in touch.® (Garrison Keillor)  I want some more patience. RIGHT NOW!

Tom M
Tom M
Joined: 2 Feb 06
Posts: 6475
Credit: 9603371617
RAC: 5781203

Read a rumor that Amd might

Read a rumor that Amd might emulate the BIG.little power savings using power limits on selected cores.

I am tickled by the idea because it basically would be a change to the bios code rather than having to tape out an entirely new chiplet and packaging.

If I am understanding the rumor it could be similar to applying "Eco-mode" but to a subset of the cores rather than the whole CPU.

I think the other thing that was mentioned is this might start in the low powered, mobile chips. But that could be just me being hazy.

Tom M

A Proud member of the O.F.A.  (Old Farts Association).  Be well, do good work, and keep in touch.® (Garrison Keillor)  I want some more patience. RIGHT NOW!

Tom M
Tom M
Joined: 2 Feb 06
Posts: 6475
Credit: 9603371617
RAC: 5781203

https://www.anandtech.com/sho

https://www.anandtech.com/show/18798/synopsysai-aipowered-eda-suite-accelerates-chip-design-and-cuts-costs-

I wonder what kind of cost savings and performance improvements you might get if you applied this to an older/larger node but used the newer micro-architecture?

Something like what Intel did for transitional CPU updates while they were stuck with older/larger manufacturing nodes.

Tom M

A Proud member of the O.F.A.  (Old Farts Association).  Be well, do good work, and keep in touch.® (Garrison Keillor)  I want some more patience. RIGHT NOW!

Tom M
Tom M
Joined: 2 Feb 06
Posts: 6475
Credit: 9603371617
RAC: 5781203

https://www.tomshardware.com/

https://www.tomshardware.com/news/amd-7950x3d-now-629-at-ebay

A Proud member of the O.F.A.  (Old Farts Association).  Be well, do good work, and keep in touch.® (Garrison Keillor)  I want some more patience. RIGHT NOW!

Tom M
Tom M
Joined: 2 Feb 06
Posts: 6475
Credit: 9603371617
RAC: 5781203

https://wccftech.com/intel-x8

https://wccftech.com/intel-x86s-paves-way-for-64-bit-only-architectures-removing-legacy-modes-simplifying-design/

A Proud member of the O.F.A.  (Old Farts Association).  Be well, do good work, and keep in touch.® (Garrison Keillor)  I want some more patience. RIGHT NOW!

Tom M
Tom M
Joined: 2 Feb 06
Posts: 6475
Credit: 9603371617
RAC: 5781203

I am beginning to wonder

I am beginning to wonder about Server cache versus consumer MB cache.  I believe my Epyc CPU cache is significantly larger than the consumer CPU cache.

Could this have implications on how high a CPU load a boinc system can sustain without becoming unresponsive?

Previous consumer CPU experience on 8c/16t and up consumer CPUs seems to indicate you need 2-4 threads not devoted to boinc projects for the system to remain responsive.

Would this be different with a server CPU with much higher CPU cache?

The reason to explore this is for cpu-based projects that could benefit from higher core/thread counts?

Tom M

 

A Proud member of the O.F.A.  (Old Farts Association).  Be well, do good work, and keep in touch.® (Garrison Keillor)  I want some more patience. RIGHT NOW!

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4979
Credit: 18792002613
RAC: 7606051

Don't understand the worry or

Don't understand the worry or supposition.

The Zen 3 core complex is exactly the same in the Zen 3 cpu desktop chips like Ryzen as it is in the Zen 3 Epyc server cpu processors.

The only difference is the number of core complexes in the package.

Same thing is true for the Zen 4 desktop cpus and the Zen 4 server cpus.

The only thing that scales larger in the Epyc cpus is the number of cores and the total L3 amount in the entire package.

 

Tom M
Tom M
Joined: 2 Feb 06
Posts: 6475
Credit: 9603371617
RAC: 5781203

What are the advantages of an

What are the advantages of an AMD EPYC 7V12 over a 7742?

Ditto 7B12 over a 7v12?

I think for the 7B12 it is the 280 watt Power Limit.  But 7v12 seems to have the same PL as a 7742.

Tom M

A Proud member of the O.F.A.  (Old Farts Association).  Be well, do good work, and keep in touch.® (Garrison Keillor)  I want some more patience. RIGHT NOW!

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4979
Credit: 18792002613
RAC: 7606051

Tom M wrote: What are the

Tom M wrote:

What are the advantages of an AMD EPYC 7V12 over a 7742?

Ditto 7B12 over a 7v12?

I think for the 7B12 it is the 280 watt Power Limit.  But 7v12 seems to have the same PL as a 7742.

Tom M

Each of those cpus has differently binned dies.  Even with the same package power, they are faster or slower than another sku.

Check the base clocks and boost clocks.  The better binned die skus have higher clocks than other skus of the same power limit.

Some of the OEM parts are deliberately lower performance at the same power limit.  A deliberate choice by the OEM to run within a lower power budget in a datacenter. Or got a better deal on price for taking the lower binned dies off the hands of the manufacturer.

7B12- base clock 2.2Ghz boost clock 3.3Ghz power limit 240W

7V12- base clock 2.0Ghz boost clock 3.3Ghz power limit 240W

7742- base clock 2.25Ghz boost clock 3.4Ghz power limit 225W

 

Filipe
Filipe
Joined: 10 Mar 05
Posts: 186
Credit: 408267800
RAC: 307873

Intel 13700 Windows 11 By

Intel 13700 Windows 11

By default it seems that boinc only run on the E-cores (windows 11), and the P-cores remains idle.

Is there a way to force boinc to use all cores (p-cores + e-cores togheter) without throttling? and without having to disable the E-cores on the bios?

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.