Multicore/thread tester - page 2

 
Alain Verleyen:

That's a known issue with current Build 2280.

Metaquotes CEO also said they will work on it.


Alain

Thank you for your prompt and clear reply

I am actually setting up 3 additional dual CPU XEON servers, all with 8x Core HT CPUs - so I was expecting to run an additional 8x2x2x3 agents
Each server has 64GB RAM, as I expected to run 30-32 agents per server.

So although MT5 says 'logical cores' v2280 actually only seems to see / run physical cores?

Is there an earlier version, or a custom version that allows all logical cores?

My 2nd 'machine', mentioned above, runs 2x 8x core HT Xeons and I enjoy all 32x cores - with exceptional processing speeds - I bought the other servers to do tasks even faster.

So, it seems my options are:

Find and use a 'HT' equipped (older) METATESTER?
Run a full MT Optimiser on each server in standalone mode (untidy)?
Wait for MQ to release a new version with HT enabled again?

 
Stephen Njuki:
Also make sure you have enough virtual memory per core/thread. Typically 2GB per thread is a minimum.

hello, you mean if buy intel Xeon 2660 (6 cores , 12 threads) i should have 32GB RAM?

 

David,

I found a quick and simple solution for this:

copy the installation folder of MetaTrader5 and run MetaTester64.exe from there to add a second batch of agents so that you can utilize a second set of agents on the same physical cores and occupy all logical processors (which could make great sense when using your own dedicated machines to run MT and have plenty of resources available in terms of RAM, etc).



David John Thompson:

Alain

Thank you for your prompt and clear reply

I am actually setting up 3 additional dual CPU XEON servers, all with 8x Core HT CPUs - so I was expecting to run an additional 8x2x2x3 agents
Each server has 64GB RAM, as I expected to run 30-32 agents per server.

So although MT5 says 'logical cores' v2280 actually only seems to see / run physical cores?

Is there an earlier version, or a custom version that allows all logical cores?

My 2nd 'machine', mentioned above, runs 2x 8x core HT Xeons and I enjoy all 32x cores - with exceptional processing speeds - I bought the other servers to do tasks even faster.

So, it seems my options are:

Find and use a 'HT' equipped (older) METATESTER?
Run a full MT Optimiser on each server in standalone mode (untidy)?
Wait for MQ to release a new version with HT enabled again?

 
The second batch gets automatically halted. Is there any other workarounds anyone knows of? Why can't they just take the restrictions off local farms?
 
Colin Leamy #:
The second batch gets automatically halted. Is there any other workarounds anyone knows of? Why can't they just take the restrictions off local farms?Hi

Hi, Colin.
Actually, those method is working. 

What you have to do is set the TCP Ports of 2nd Batch different with 1st Batch. If the TCP Ports of 2nd Batch accidentally has set same with 1st Batch, then you must uninstall the Agent one by one (in 2nd Batch) then re-Run and re-Set the TCP Port (and click Add) for the Agent of 2nd Batch.

They will look like this:


 
I have a server computer with two Intel Xeon E5 with a total of 20 cores, 40 threads.
If I start an optimization (slow complete algorithm) with all cores (threads) on, it does not work for me.
The trick is to enable the threads per sets, I often start with 4 core threads, then the optimization starts, if that runs I add 8 threads, if that runs I add more threads, etc.
Hope this will work for you!
Reason: