• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

would anyone with a fast system be willing to do me a favor?

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Raynor, could you post your smaller project and list a recommended device that we can all use to get more accurate results?
 
Originally posted by: SpecialK
Raynor, could you post your smaller project and list a recommended device that we can all use to get more accurate results?
Click to expand...

OK, I've got a nice one that I just compiled in 2:15 minutes on my computer. Let me see if I can get the files to fit in a reasonable amount of space for hosting. I'll upload it in a minute once I get my files sorted out.
 
OK, I uploaded the file, Click me

You're going to have to do the following

1 - Open the .quartus file (in Quartus, obviously)
2 - In the panel on the right, click Files. Now right-click g30_surveyor_machine.gdf, then left click "Set Compiler Focus to Current Entity"
3 - In the toolbar on top, select Assignments -> Device
4 - Choose Family: Flex10K
5 - In Available Devices choose EPF10K70RC240-4
6 - Click OK, click the purple "Play" button in the top toolbar

On My 2.4 GHz P4C (overclocked to 3.0 GHz) my compile time is approx. 2:15 on the first run. If you run it again, there are a bunch of files that it doesn't need to generate again, so my compile time drops to ~1:45 on subsequent runs.
 
Originally posted by: PlatinumGold
if it isn't smp aware, is it better to run it with HT turned off??
Click to expand...

I don't think there's a single situation where HT lowers your performance, so you should leave it on. In fact, if one thread is dealing with the software and the other is dealing with Windows, you get a performance boost even if the software isn't SMP aware.
 
Originally posted by: RaynorWolfcastle
OK, I uploaded the file, Click me

You're going to have to do the following

1 - Open the .quartus file (in Quartus, obviously)
2 - In the panel on the right, click Files. Now right-click g30_surveyor_machine.gdf, then left click "Set Compiler Focus to Current Entity"
3 - In the toolbar on top, select Assignments -> Device
4 - Choose Family: Flex10K
5 - In Available Devices choose EPF10K70RC240-4
6 - Click OK, click the purple "Play" button in the top toolbar

On My 2.4 GHz P4C (overclocked to 3.0 GHz) my compile time is approx. 2:15 on the first run. If you run it again, there are a bunch of files that it doesn't need to generate again, so my compile time drops to ~1:45 on subsequent runs.
Click to expand...
Obviously. I am "opening" it wrong, as I open it, and there is no right pane. I un-rared it, and opened the file and ran it, and its been 10 minutes, and at 32%.
 
Link works fine...
Click to expand...

Tue May 04 15:47:38 2004 Connecting ftp.altera.com:21
Tue May 04 15:47:46 2004 Connecting ftp.altera.com [IP=66.35.227.3:21]
Tue May 04 15:47:46 2004 Socket connected, Waiting for welcome message.
Tue May 04 15:49:16 2004 Time out.
Tue May 04 15:49:16 2004 Error occured!
Tue May 04 15:49:16 2004 Wait 5 second for retry
Tue May 04 15:49:21 2004 Connecting ftp.altera.com [IP=66.35.227.3:21]
Tue May 04 15:49:21 2004 Socket connected, Waiting for welcome message.
 
ok, here's the deal. i went in, disable HT and guess what, the program is SMP capable, but only with multi processors not with HT.

i'll post final times for my dual Xeon 2.8 mhz, 1 ghz RDRam system later.
 
Originally posted by: RaynorWolfcastle

On My 2.4 GHz P4C (overclocked to 3.0 GHz) my compile time is approx. 2:15 on the first run. If you run it again, there are a bunch of files that it doesn't need to generate again, so my compile time drops to ~1:45 on subsequent runs.
Click to expand...

I get 1 minute 54 seconds for the first run, Athlon 3000 @ stock (1Gb PC2700), 1 minute 47 seconds for the second run.

<shrug>

Here's proof, as I thought you'd ask 🙂

Linky
 
Originally posted by: jagec
Originally posted by: PlatinumGold
if it isn't smp aware, is it better to run it with HT turned off??
Click to expand...

I don't think there's a single situation where HT lowers your performance, so you should leave it on. In fact, if one thread is dealing with the software and the other is dealing with Windows, you get a performance boost even if the software isn't SMP aware.
Click to expand...

HT Enabled, it compiled it in 51 seconds. HT Disabled it compiled in 42 seconds.

btw, i expaned the .rar file to two different folders to make sure that one wasn't affected by the previous compilation. AND i ran it first with HT Disabled.

so, my conclusion, on this particular program which apparently ISN'T HT aware BUT is smp capable i get better performance, when nothing is running in the background with HT Off. IF i had a lot of stuff in the background i suspect it could actually slow down more with HT off . . .
 
With my P4c 3.4GHz with 1GB, it took 35 seconds for the first run and 31 seconds in subsequent runs. Decided to run 2 instances of the program, it took 49 seconds for both to run. With one instance of the program but running Seti in the background, it took 41 seconds to run.

update

disabled HT

deleted data files and started with fresh files

32 seconds for 1st run and 29 seconds in subsequent runs
2 instances running, 1st one completed in 42 seconds and the 2nd one in 56 seconds
and here is now the odd one, seti running in background, it took 30 seconds to run one instance of the program
 
Originally posted by: eelw
With my P4c 3.4GHz with 1GB, it took 35 seconds for the first run and 31 seconds in subsequent runs. Decided to run 2 instances of the program, it took 49 seconds for both to run. With one instance of the program but running Seti in the background, it took 41 seconds to run.
Click to expand...

That's impossible unless it didn't complete successfully. The rest of the scores are inline with what they should be.

PlatinumGold, it's impossible for it to be SMP capable but not HT-aware. HT-awareness just means that it's multi-threaded which is the same as SMP-capable. What is possible, however, is that there are many threads that use the same part of the CPU (ALU unit, in this case) in which case HT would hurt you since all threads want to use the same part of the CPU while context changes aren't free, performance-wise.

Markfw900. The panel on the right contains information on the current project. Now that I think about it, I may have moved to to the right myself, it may be on the left for you.

Edit: Oh yeah, as long as everything completes successfully, don't worry about the warnings.
 
Originally posted by: RaynorWolfcastle
Originally posted by: eelw
With my P4c 3.4GHz with 1GB, it took 35 seconds for the first run and 31 seconds in subsequent runs. Decided to run 2 instances of the program, it took 49 seconds for both to run. With one instance of the program but running Seti in the background, it took 41 seconds to run.
Click to expand...

That's impossible unless it didn't complete successfully. The rest of the scores are inline with what they should be.

PlatinumGold, it's impossible for it to be SMP capable but not HT-aware. HT-awareness just means that it's multi-threaded which is the same as SMP-capable. What is possible, however, is that there are many threads that use the same part of the CPU (ALU unit, in this case) in which case HT would hurt you since all threads want to use the same part of the CPU while context changes aren't free, performance-wise.

Markfw900. The panel on the right contains information on the current project. Now that I think about it, I may have moved to to the right myself, it may be on the left for you.

Edit: Oh yeah, as long as everything completes successfully, don't worry about the warnings.
Click to expand...



when i open task manager and look at CPU Usage with HT enabled, only one of the 4 windows (i have dual processors each with 2 ht instances) show activity. with HT Disabled, open Task Manager and look at CPU Usage, both cpu windows show about even usage.

based on that, i concluded that it was smp aware (could use 2 processors) but not HT aware, i could be wrong because if i were right and it was using both CPU's i should get faster results than EELW.

so, mb it isn't smp aware, but it does seem to show even distribution of burden between the two cpus but neither is running 100%.
 
for God's sake man..just buy a new computer...
i'm going to make a bold prediction....
if you get a 3.0 Ghz machine....it's going to run at least twice as fast...
 
Originally posted by: RaynorWolfcastle
Originally posted by: eelw
With my P4c 3.4GHz with 1GB, it took 35 seconds for the first run and 31 seconds in subsequent runs. Decided to run 2 instances of the program, it took 49 seconds for both to run. With one instance of the program but running Seti in the background, it took 41 seconds to run.
Click to expand...

That's impossible unless it didn't complete successfully. The rest of the scores are inline with what they should be.
Click to expand...

What do you mean "That's impossible"???
 
Originally posted by: RaynorWolfcastle
I mean that it's impossible for your 3.4C beats my 3.0C by a factor of more than three. Doesn't that seem a little weird to you?
Click to expand...

you re right. i went back and checked everyones numbers, Eelw's numbers are out of line. Quartus IS smp enabled. i think your analysis is correct, i also think that with a dual system like mine, disabling HT is better for extremely cpu intensive tasks. for normal use, it is probably better HT enabled.
 
Originally posted by: RaynorWolfcastle
I mean that it's impossible that your 3.4C beats my 3.0C by a factor of more than three. Doesn't that seem a little weird to you?
Click to expand...

Good point. I followed the steps that you provided. The only difference was step 2. When I right-click on the g30xxxx file, there was no option to "Set Compiler....." - just Open/Remove/Set as Top Level Entity/Properties/close. So I skippped step 2. Would this really make that much a difference in run time?
 
Yes, it makes a big difference, you're synthesizing one of the components of the device instead of compiling the entire device. I'm not sure what went wrong, but the third option in the compiler should be to "Set Compiler Focus to Current Entity".

In any event, it's really not that important if you can't get a decent bench on a very specific program CAD program that's used only by a few CEs/EEs. Especially, since your benchmark is some ghetto circuit I designed for one of my classes that serves no practical purpose. 😉
 
EDIT: OK, on my machine, it took 3:41, thats more reasonable (Tbird 1.2 GHz)

On another note, the original large design had a serious issue with it that was causing it to take forever to compile. One of the files in the design was provided to me and apparently Quartus didn't like it. I replaced that file with another one that came with Quartus (made by Altera, same basic functionality) and suddenly the compile time for my original design dropped to about 3 min. Maybe I don't need a new PC after all.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top