Hi Ben,

Thanks for your reply.

I'd always assumed that Xeon's were faster because they had a different internal architecture to the Core range. I guess the heat and power constraints of mobiles mostly negate those differences.

Thanks again, Alistair

On Tue, 24 Sep 2019 at 12:41, Ben Coman btc@openinworld.com wrote:

On Tue, 24 Sep 2019 at 13:55, Alistair Grant akgrant0710@gmail.com wrote:

...

I hope you don't mind a somewhat off-topic question: Does anyone happen to know which will be faster for running Smalltalk (i.e. single threaded application, and given the rest of the laptop is the same):

Intel Core i9-9880H vs. Xeon E-2276M

Intel® Xeon E-2276M, 6 Core, 12M Cache, 2.80GHz, 4.70GHz Turbo, 35W, vPro

Intel® CoreTM i9-9880H Processor, 8 Core, 16MB Cache, 2.30GHz, 4.80GHz Turbo, 35W, vPro

Purely guesswork... that the 12MB cache is already huge and the 16MB not much more impactful. Obviously we ignore the number of cores. Clock frequency should have the biggest impact for single-threaded apps, which primarily points to the Xeon, but I don't know how to judge the i9 turbo being faster. Probably on a laptop you won't have the thermal overhead to run turbo too much.

By the numbers...

PassMark - CPU Mark - Single Thread Performance Intel Xeon E-2276M @ 2.80GHz 2,684 (78%) $450.00 Intel Core i9-9880H @ 2.30GHz 2,548 (74%) $556.00

https://www.cpubenchmark.net/singleThread.html https://www.cpubenchmark.net/compare/Intel-i9-9880H-vs-Intel-Xeon-E-2276M/34...

cheers -ben

Xeon's also get to take advantage of ECC RAM and i9's (and the other i series) don't.

https://perspectives.mvdirona.com/2009/10/you-really-do-need-ecc-memory/

Ben Coman wrote

On Tue, 24 Sep 2019 at 21:12, Alistair Grant <

akgrant0710@

> wrote:

...

Hi Ben,

Thanks for your reply.

I'd always assumed that Xeon's were faster because they had a different internal architecture to the Core range.

No, that would cost more. The distinction between Xeon/i9/i7/i5 is simply how well they test after manufacture.

Xeons were premium because there had to be less manufacturing errors to end up with larger caches. But your i9 there has a larger cache, so maybe that quality dimension has shifted. This is interesting reading... https://www.quora.com/How-does-Intel-design-and-produce-so-many-models-of-CP...

cheers -ben

I guess the heat

...

and power constraints of mobiles mostly negate those differences.

Thanks again, Alistair

On Tue, 24 Sep 2019 at 12:41, Ben Coman <

btc@

> wrote:

...

...

On Tue, 24 Sep 2019 at 13:55, Alistair Grant <

akgrant0710@

>

...

wrote:

...

...

I hope you don't mind a somewhat off-topic question: Does anyone happen to know which will be faster for running Smalltalk (i.e. single threaded application, and given the rest of the laptop is the same):

Intel Core i9-9880H vs. Xeon E-2276M

Intel® Xeon E-2276M, 6 Core, 12M Cache, 2.80GHz, 4.70GHz Turbo, 35W,

vPro

...

...

Intel® CoreTM i9-9880H Processor, 8 Core, 16MB Cache, 2.30GHz, 4.80GHz Turbo, 35W, vPro

Purely guesswork... that the 12MB cache is already huge and the 16MB

not much more impactful.

...

Obviously we ignore the number of cores. Clock frequency should have the biggest impact for single-threaded

apps,

...

which primarily points to the Xeon, but I don't know how to judge the

i9 turbo being faster.

...

Probably on a laptop you won't have the thermal overhead to run turbo

too much.

...

By the numbers...

PassMark - CPU Mark - Single Thread Performance Intel Xeon E-2276M @ 2.80GHz 2,684 (78%) $450.00 Intel Core i9-9880H @ 2.30GHz 2,548 (74%) $556.00

https://www.cpubenchmark.net/singleThread.html

https://www.cpubenchmark.net/compare/Intel-i9-9880H-vs-Intel-Xeon-E-2276M/34...

...

cheers -ben

-- Sent from: http://forum.world.st/Squeak-VM-f104410.html

Hi Paul & Stephan,

On Tue, 24 Sep 2019 at 17:26, Stephan Eggermont stephan@stack.nl wrote:

...

Op 24 sep. 2019 om 17:17 heeft Paul DeBruicker pdebruic@gmail.com het volgende geschreven:

Xeon's also get to take advantage of ECC RAM and i9's (and the other i series) don't.

On a laptop? Unlikely. The actual speed difference is probably only dependent on the quality of the cooling and allowed temperatures. With the first 8-core MBPs there were complaints that they were not actually faster because of the heat limitations

Right, ECC is supported on Dell's 3541 with a Xeon processor: https://www.dell.com/en-us/work/shop/workstations-isv-certified-dell/precisi...

Thanks, Alistair

Hi Ben,

On Tue, 24 Sep 2019 at 16:39, Ben Coman btc@openinworld.com wrote:

On Tue, 24 Sep 2019 at 21:12, Alistair Grant akgrant0710@gmail.com wrote:

...

Hi Ben,

Thanks for your reply.

I'd always assumed that Xeon's were faster because they had a different internal architecture to the Core range.

No, that would cost more. The distinction between Xeon/i9/i7/i5 is simply how well they test after manufacture.

Xeons were premium because there had to be less manufacturing errors to end up with larger caches. But your i9 there has a larger cache, so maybe that quality dimension has shifted. This is interesting reading... https://www.quora.com/How-does-Intel-design-and-produce-so-many-models-of-CP...

Now that you mention this, I do remember reading something similar (a long time ago).

Thanks! Alistair