I want to build a PC to work with MT5, what do you advise and why? - page 66

[Aleksander]

Well and for mt5, to speed up testing, you can prepare ticks, removing duplicates and id, reducing the time of running the test in tens of times

[lynxntech]

And do not compare different platforms, northern 4 channels and consumer 2, and different manufacturers AMD has its own route paths from ram to cpu, longer, but there is additional caching, in general they can not be compared.

small monster TA140, changed to a medium Dark Rock 4, and there is no difference)))) minus almost 5000rub.

[lynxntech]

Does the 3D cache actually make a difference in the tests? Has anyone run any benchmarks?

[Michael Charles Schefe]

lynxntech #:
Does the 3D cache actually make a difference in the tests? Has anyone run any benchmarks?

i have not done any recent testing of 3d cache-ing, but even back then -- the same testing with 3d cache varied greatly between system to system, intel to amd, and even when transfering the gfx card between the 2 systems. so i think that you can not make any generalised assumptions from any results between any brand or type of parts in each system.

[Gio Rendel Masagca Rivadillo]

For MT5 backtesting and Strategy Tester, prioritize a high-core CPU like a Ryzen 9 9900X or better, paired with at least 32–64GB RAM and a fast NVMe SSD since MT5 performance is mostly CPU-bound and scales with cores during optimization. The more the cores the better for faster accurate backtesting 

[lynxntech]

Michael Charles Schefe #:
I haven’t carried out any 3D cache testing recently, but even back then, the results of the same tests with 3D cache varied significantly from system to system, from Intel to AMD, and even when swapping a graphics card between two systems. I therefore believe that it is not possible to draw any general conclusions based on results obtained for components of different brands or types within each specific system.

I don’t mean on the graphics card,

Processors

AMD 3D V-Cache™ Technology

https://www.amd.com/en/products/processors/technologies/3d-v-cache.html

I’m interested in the Ryzen 9000 3D V-Cache series. Has anyone swapped a standard model for an X3D model?

[Aleksey Vyazmikin]

lynxntech #:

I don’t mean on the graphics card,

Processors

AMD 3D V-Cache™ Technology

https://www.amd.com/en/products/processors/technologies/3d-v-cache.html

I’m interested in the Ryzen 9000 3D V-Cache series. Has anyone swapped a standard model for an X3D model?

There was a report that this cache is on a single chiplet and only operates quickly with the cores on that chiplet; and since not all cores are used in games – that is, there’s a boost – whilst the speed of cache requests from the other die is significantly lower, which ultimately offers no advantages. Based on this logic, this additional cache would serve no purpose in optimising across all cores.

[Michael Charles Schefe]

lynxntech #:

I don’t mean on the graphics card,

Processors

AMD 3D V-Cache™ Technology

https://www.amd.com/en/products/processors/technologies/3d-v-cache.html

I’m interested in the Ryzen 9000 3D V-Cache series. Has anyone swapped a standard model for an X3D model?

if you trust any of the youtubers, then, some of those show that they hold the higher ghz/mhz for a bit longer on the first 2-4 cores, however it only happens on some games, not all, and most certainly not noticable difference in normal system operation, unless you do count the extra heat that you may or may not notice.

[lynxntech]

Aleksey Vyazmikin #:
There was a report that this cache is located on a single chiplet and only operates quickly with cores on that chiplet; and since not all cores are used in games—that is, there is a boost— whilst the speed of cache requests from another chiplet is significantly lower, which ultimately offers no advantages. Based on this logic, optimising across all cores would not benefit from this additional cache.

In this regard, a model has recently been released which, on two die

if some of the cached cores finish earlier, that’s very good

[lynxntech]

Michael Charles Schefe #:
If you are to believe certain YouTubers, some of them demonstrate that the first 2–4 cores maintain higher GHz/MHz values for slightly longer, however, this only happens in some games, not all of them, and the difference is certainly not noticeable during normal system operation, apart from the extra heat, which you may or may not notice.

After all, our setup is rather unusual.) There’s no point watching videos about this on YouTube