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
CPU: Athlon 64 X2 3800+ (2000 MHz), cache 2x512 KB L2 (Socket 939) RAM: 4*512MB PC3200
Script test 90 sec, optimization 318 sec, ticks 1:09:06
If I place terminal in RAM, results will be slightly different )
Script test 86s, optimization 297s, ticks 49:01Funny :)
Download the latest InfInstaller for 9xx,3,4 Intel chipset and make sure (but you'll have to dig around in inf files), it's a bit unclear what north bridge they will be compatible with and what will be the name of the stones themselves (i3/i4). (So they'll remove the north bridge from the stone, they will...)
Exactly, smiled, especially the last sentence. Not only will they not remove anything, they will also add integrated graphics.
Why to invent when regularly occurring "leaks" outline very clearly all perspectives of Nehalem's development?
And why to spy on inf-files? There is nothing "criminal" about Nehalem LGA775 release. The statements are not unfounded - I've downloaded and checked them. If suddenly you think that presence of NehalMEX.inf file says about allegedly forthcoming Nehalem LGA775 release - that's an obvious misconception.
Or better yet, tell me exactly which files (or even lines in them) made you come up with such fantastic ideas - let it make more than just me smile.
2 All - sorry for the off-topic. I'm no expert in trading or mathematics, but I'm good with the matchbook. And technically illiterate statements, and even more so fiction like the one described above, leads to a certain loss of self-control ...
Entered the results of TorBar and Imp120 in the table.
How can we arrange for a link to the tests and report form to be displayed at all times? It is difficult for those wishing to join the test to hunt through the whole branch. And the requirements/explanations for the test were then added. Maybe on people.ru to make a page? Or just upload an archive with tests and description on file sharing?
Any suggestions?
Mathemat, add mine as well:
script:
Optimisation 2:54.
My processor:
Memory:
Had a closer look at the table. There are actually some oddities. And not only with four2one, but also with begemot61.
Both are quite predictable results when running a script - P4 is very much behind any modern CPU in efficiency, Opteron is on the level of Core 2, Xeon is slightly better than Core 2.
But when optimizing, "miracles" begin - Athlon 64 X2 5050e is 2.3 times faster than Athlon 64 X2 3800+ with only 1.3 times the frequency difference. The Pentium 4 670 also becomes one of the fastest processors. The Opteron 2439 SE beats the much higher frequency Phenom II X3 720 @ 3.72 GHz, and the Xeon even gets out of 100 seconds...
I'm guessing there's a methodological error somewhere. The fundamental difference between the script and the Expert Advisor is that the latter depends not only on history of quotes but also on properties of the trading account. Therefore, in order to verify the correctness of the results of speed measurements it is highly recommended to review the results of the tester run (at arbitrarily selected parameters of the Expert Advisor) - tab "report". And compare it with "slower" systems.
A description of the optimisation test.
We take the Moving Average Expert Advisor from the standard delivery of MT. Save the existing EURUSD1 history by exporting it. Increment the history on the minutes. Make sure that the history covers the testing period! Make sure that there are no files in the tester folder \tester\history. If there are - delete them! Then you set these parameters in the tester:
Parameters in the window below can be loaded from t.set file or set manually.
Remove genetic algorithm flag!!! It is clear why.
That's it. The report will be the first screen with the optimization time.
Test description on the script.
Everything is simple - just run the script and take a screenshot of the alert window.
Necessary information about the computer.
Two screenshots of CPU-Z utility tabs: CPU and RAM.
Everest report is also welcome. Of course, as an attached file.
Archive contents: optimization setup file t.set, script file, test description file, latest CPU-Z version. I've embedded the archive to this address http://files.mail.ru/Z5PWGJ. Now you can just give a link to it.
OK, you can't edit the table on page 28 now, three days have passed. OK, moving it here, adding the Docent results:
joo
Athlon 64 X2 3800+ (2000 MHz), cache 2x512 KB L2
DDR2 PC-5360 2GB
82.07*2(?)=164.14
310*2=620
Mathemat
Core 2 Duo E7200@2.53, cache 3 MB L2
4GB RAM PC-6400
46.27*2.53=117.06
213*2.53=538.89
Svinozavr
Celeron 900 @ 2.20 GHz, cache 1 MB L2
DDR2 PC-6400 2GB
52.18*2.2=114.8
206*2.2=453.2
benik
Celeron 325 @ 2.53 GHz, cache 256 K L2
DDR2 PC-6400 1GB
103.3*2.53=261.35
438*2.53=1108.14
begemot61
Pentium 4 670 @ 3.8 GHz, cache 2 MB L2
DDR2 PC-4266 2GB
78.57*3.8=298.57
169*3.8=642.2
BLACK_BOX
Athlon 64 X2 4200+ @ 2.2 GHz, cache 2x512 KB L2
DDR1 PC-3200 (?) 3 GB
77.84*2.2=171.25
forex-k
Core 2 Duo Q8200 @ 2.33 GHz, cache 2x2 MB L2
RAM 4 GB PC-6400
46.84*2.33=109.14
189*2.33=440.37
Phenom II X3 720 @ 3.72 GHz, cache 3x512 KB L2 + 6 MB L3
four2one
Athlon 64 X2 5050e @ 2.6 GHz, cache 2x512 KB L2
RAM 4 (8) GB PC-5970
60*2.6=156.0
134*2.6=348.4
begemot61
6-Core Opteron 2439 SE @ 2.8 GHz, cache 6x512 KB L2 + 6 MB L3
DDR2 4 GB PC-5333
42.33*2.8=118.52
95*2.8=266
begemot61
Xeon W5590 @ 3.47 GHz, cache 4x256 KB L2 + 8 MB L3
DDR3 PC-10670 12GB
27,53*3,47=95.53
62*3.47=215.14
Dmido
Pentium 4 @ 3 GHz, cache 512 KB L2
DDR1 PC-3200 1.15 GB
64.49*3=193.47
315*3 = 945
TorBar
Intel Celeron 331 @ 2.66 GHz, cache 256 KB L2
DDR1 (PC-3200) 1.5GB
105.49*2.66=280.60
386*2.66=1026.76
imp120
Athlon 64 X2 3800+ @ 2 GHz, cache 2x512 KB L2
DDR1(?) PC-3200 2 GB
90*2.0=180
318*2.0=636
Docent
Core 2 Duo E6550 @ 3 GHz, cache 4 MB L2
DDR2 PC-6864 2 GB
40.35*3.0=121.05
174*3.0=522
imp120
Mobile Core 2 Duo P8600 @ 2.4 GHz, cache 3 MB L2
DDR2 4GB PC-6400
44.99*2.4=107.98
201*2.4=482.4
Vinin
Core 2 Duo E8400 @ 3.00 GHz, cache 6 MB L2
DDR2 4GB PC-6400
36.99*3.0=110.97
152*3.0=456
OK, you can't edit the table on page 28 now, three days have passed. OK, move it here, adding Docent's results:
You have an error in your own optimization time of 213*2.53=538.89.
Your run time was 2:49, i.e. 169 sec, not 213. Where did the 213 come from?
Mathemat, add mine as well:
script:
Optimisation 2:54.
My processor:
Memory:
Had a closer look at the table. There are actually some oddities. And not only with four2one, but also with begemot61.
Both are quite predictable results when running a script - P4 is very much behind any modern CPU in efficiency, Opteron is on the level of Core 2, Xeon is slightly better than Core 2.
But when optimizing, "miracles" begin - Athlon 64 X2 5050e is 2.3 times faster than Athlon 64 X2 3800+ with a frequency difference of only 1.3 times. The Pentium 4 670 also becomes one of the fastest processors. The Opteron 2439 SE beats the much higher frequency Phenom II X3 720 @ 3.72 GHz, and the Xeon even gets out of 100 seconds...
I'm guessing there's a methodological error somewhere. The fundamental difference between the script and the Expert Advisor is that the latter depends not only on history of quotes but also on properties of the trading account. Therefore, in order to verify the correctness of the results of speed measurements it is highly recommended to review the results of the tester run (at arbitrarily selected parameters of the Expert Advisor) - tab "report". And compare it to "slower" systems.
I had my doubts too. With the script, everything seems to be correct. What about the EA? Maybe we should run it once with default settings before optimization. At the beginning of trades it will be clear that the history is correct. Well, according to the time of losing or the number of deals, it will work correctly and the quotes history is adequate.
I had my doubts too. With the script, everything seems to be correct. What about the EA? Maybe we should run it once with default settings before optimization. At the beginning of trades it will be clear that the history is correct. Well, according to the sinking time or amount of trades, the trading robot works correctly and the quotes history is adequate.
Quotes may be quite adequate, while properties of a trading account may be different to such an extent that it will affect trades. Therefore it would be desirable to look (at arbitrary, but published by you settings) on "report" tab of tester, in test mode, but not optimization.