AMD or Intel as well as the memory brand - page 17
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
The trend is clear. Soon we'll be reaching the point where we're talking about a tube-based computer. Core Voltage 150V.
As long as it's not an armoured vehicle computer.
)))
Here's a test of a budget 2-core AMD build from 3 years ago.
Colleagues from AMD camp, join in!
Show something really manly from AMD that will tear Intel on maths on this script (MT4, of course)! There's been a lot of word of mouth, but the tests so far are not impressive. Just a reminder, the best result so far is 46.
Skymaster, where is your test? The point is not even what is "better" - AMD or Intel - but to see what each platform can do in a real, concrete MT4 terminal and in this particular script.
Lye! I don't understand anything at all! In order to equalise the weights of each test to one figure, e.g. 100, I tried increasing the number of cycles in the first test, hoping that the execution time depends linearly on the number of cycles. And what happened? The time increased by an order of magnitude in other tests too!
I tried to declare an array before the third test, I looked into the Task Manager and couldn't believe my eyes - memory for the array had been allocated even before the array was declared! Anyone have anything to say about this?
Andrei, let's keep the test the same for now. There are already some statistics, after all.
It's simpler than that maybe. You have an overclocked - heating - trolling stone. I.e. idle strokes are inserted. The frequency doesn't change. Yes. It's very possible. Try to measure the temperature and then reduce the frequency to nominal.
Andrei, let's keep the test the same for now. There's already some statistics, though.
No no, I'm not going to bolamut, I'm just puzzled by the very strange behaviour of my hardware (or maybe the MT compiler?)
And one more thing. Most likely, I'm 1000% sure, the code is compiled for the Intel platform and AMD stones won't be able to show anything great and manly in this test.
Conclusion: Serious MT math calculations should preferably be done on Intel. I think only the developers of the terminal will be able to convince me otherwise.
Of course it would be interesting to see the test on a serious modern AMD stone. Who has such a stone?
And one more thing. Most likely, I'm 1000% sure of it, the code is compiled for Intel platform and AMD stones won't be able to show anything big and manly in this test..,
They have the same set of commands - x86 + all that stuff. The way of queuing, decoding, mixing for filling in empty clock cycles and so on and execution itself are different. Many things may be done differently. The cache is the same at the same volume.
So I would not say that compilation is made for a specific platform, especially since the compilation result is not a set of instructions for a stone but the so called byte-code as for java-machine or .NET.
Although, of course, everything is possible)))
Mathemat писал(а) >>
the best result so far is 46.