AMD or Intel as well as the memory brand - page 66

 
kombat >> :

More on the highlighted one, please...


Yes discussed here already. You need to do some research on HOW exactly the tester uses resources. That's quite a task! ))) And then write an emulator. Yeah... I wouldn't do it...

Why do it with a live tester is unclear. It's not like it's a car that a live tester would choke to death.

 
Docent >> :

And by the way, what's the "hitch" and how does SSE2 help there? And where did this information come from in the first place? Can you give me the link?

That's my personal opinion...

I got it from wikipedia.

If it's used, it must be advantageous in some way...


By the way, I apologize for my mistakes about doubles.


SSE( Streaming SIMD Extensions) is a SIMD( Single Instruction, Multiple Data) instruction set developed by Intel and first introduced in Pentium III series processors as a response to a similar instruction set 3DNow! from AMD which was introduced a year earlier. The original name of these instructions was KNI, which stands for Katmai New Instructions (Katmai was the name of the first version of the Pentium III processor core).

SSE technology allowed to overcome 2 major problems of MMX - when using MMX, it was impossible to use co-processor instructions simultaneously, because its registers were used for MMX and real number operations.

SSE includes in processor architecture eight 128-bit registers(xmm0 to xmm7) each of which is treated as 4 sequential single-precision floating-point values. SSE includes an instruction set which performs operations on scalar and packed data types.

A performance advantage is gained when the same sequence of operations on different data needs to be performed.

SIMD blocks are implemented by paralleling computational process between data, i.e., when many data streams pass through one block in turn.

и


SSE2( Streaming SIMD Extensions 2) is a SIMD(Single Instruction, Multiple Data) instruction set developed by Intel and first introduced in Pentium 4 series processors.

SSE2 uses eight 128-bit registers (xmm0 to xmm7), included in x86 architecture with SSE extension input, each of which is treated as 2 double precision sequential floating point values. SSE2 includes an instruction set which performs operations on scalar and packed data types. SSE2 also includes instructions for integer-streaming in the same 128-bit xmm registers, making this expansion preferable to the MMX instruction set, which was introduced much earlier.

A performance advantage is gained when the same sequence of operations is to be performed on a large set of identical data.

 
Svinozavr >> :

Yes, it has already been discussed here. For this purpose it is necessary to make a research work HOW exactly the tester uses resources. That's quite a task! ))) And then write an emulator. Yeah... I wouldn't do it...

Why do it with a live tester is unclear. It's not a car that you'd want to tear apart in a live one.

The complicated stuff doesn't need to be built there, of course...

But for, say, a primary, "cold sight tube", it could be hardwired.

And in case of obvious inconsistency to run the optimizer in live... >> for the purity of the experiment.

 

kombat, the question was not about what SEx is, but how it is used in MT5 and where the information comes from.

 
Docent >> :

kombat, the question was not about what SEx is, but how it is used in MT5 and where the information comes from that is used at all.

I have no idea. ;)

In a week the public beta-test will start, there and we will know.

In the meantime, we're using 4 and everything for it...

 
Docent >> :

kombat, the question was not about what SEx is, but how it is used in MT5 and where the information comes from.

The 'MQL5 wishes'.

>> 'History of MQL4?

 

Thank you!!!

 

I have a (clean XP32) running in a virtual machine - hopefully this will not affect the tests too much

i.e. on top of CentOS 64 / only vmware 6.5.1 126130 is running during the tests

all stopped - i.e. excluded almost all other processes



hardware

I7 920 12gb DDR3

the virtual machine has allocated 3072 MB for this copy


these runs were just following - didn't change anything

probably got cached.



above is with genetics

will cook without it



below is without genetics


 
YuraZ писал(а) >>

I have (pure XP32) in a virtual environment - let's hope this does not affect the tests too much

i.e. on top of CentOS 64 / only vmware 6.5.1 126130 is running for the duration of the test

all stopped - i.e. virtually all other processes

hardware

I7 920 12gb DDR3

virtual machine has allocated for this copy 3072 mb

these runs were just following - didn't change anything

probably cached

Yura wanted to see your results.

I did.

Now all that's left is the cost.

 

Yura! It's been stressed many times: the tester's cache and histories must be empty before testing!

Why no one reads instructions?))) Although they are hard to find here - that's right.