Too early to say, but 4 WU in 8000 secs is only 10% faster than a 390X running 1 at a time, far away from the promissed bost on performance, something else must be happening.

Maybe 4 WU at a time is too much due the I5 CPU who actualy feed the GPU. Who knows.

Hah, I see you beat me here. His i5 is strong (6600), with good single-thread performance. I don't think it is the CPU. Here is the task list:

https://einstein.phys.uwm.edu/results.php?hostid=12275081&offset=0&show_names=0&state=0&appid=29

BRP6 results are always fairly constant. I've asked him to check his core and memory clocks to see if the downclocking issue from Maxwell series persists.

---

Hey guys, as I mentioned in this post above, since I am getting one, and will be installing it on the 48 core system that I am configuring right now, any thoughts you might have about proper apps to use, since it is a fresh load? Won't be installing vid drivers till I get the card, so will probably have to reinstall the optimized apps if that is suggested to install today as the CPU cruncher it will start out life as, till the vid card arrives. Or just run it stock? And is there a specific version I should install now, to be ready for when the card arrives? I am open to any and all suggestions.

---

Hey guys, as I mentioned in this post above, since I am getting one, and will be installing it on the 48 core system that I am configuring right now, any thoughts you might have about proper apps to use, since it is a fresh load? Won't be installing vid drivers till I get the card, so will probably have to reinstall the optimized apps if that is suggested to install today as the CPU cruncher it will start out life as, till the vid card arrives. Or just run it stock? And is there a specific version I should install now, to be ready for when the card arrives? I am open to any and all suggestions.

I typically find the best performance with the AVX app, but with that many tasks running at once SSE3 might be more efficient. It could take running both for a while to see. Either option should be more efficient than the base CPU app.

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

Al, I would run stock first just to get a baseline.

Let it run for a few hours and then you can get an average.

After that you can install an app_config to run more than 1 at a time and see what the times are like.

After that, maybe installing the commandline into the app_config to try and get the best speed

So, is that SSE3, SSSE3, or SSSSSSSSE3? ;-) lol This is a Xeon V2 proc, I thought that SSE3 (or whatever) was for older procs, a few more gens older than this one? Or does this proc fall under that version as well? Or, is it just simpler code that might run more efficiently? Too much to know! ;-)

---

Ok, Zalster. Just to clarify, once I get BOINC installed today (sans any GPU as it is a server motherboard obviously) just run it with the stock install. No Lunatics? As to the more than 1 at a time, I presume you mean once I get the vid card installed. If I install the optimized app for the CPU before the vid card is installed, will there be any repercussions when the time comes to install the card? So much that I don't yet know...

---

So, is that SSE3, SSSE3, or SSSSSSSSE3? ;-) lol This is a Xeon V2 proc, I thought that SSE3 (or whatever) was for older procs, a few more gens older than this one? Or does this proc fall under that version as well? Or, is it just simpler code that might run more efficiently? Too much to know! ;-)

For MBv8 there are only AVX, SSE3, & SSE2 applications in the Lunatics installer 0.44. The other options are for the MBv7 app where there were more options. Logically whatever selection is made that is not one of the MBv8 options would select the next lower options for MBv8. Such as selecting SSSE3, SSE4.1, or SSE4.2 should install that version of MBv7 & the SSE3 version of MBv8.

SIMD instructions are normally additive. So if you have a SSE 4.2 CPU it supports all previous versions as well.
SSE3 might just be more efficient than AVX on your 24c/48t configuration. The highest level SIMD instruction set is not always the most efficient option.

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

Correct no Lunatics.

Since we want to see how it works without the optimized programs in it first, no lunatics

So, no lunatics. Run stock.. Get GPU... Run stock

After an average, install an app_config for you can run more than 1 stock at a time on the GPU

Figure out what the best time is.

After that, install lunatics and can get app_config to start again with 1 per card and work your way up until you find the fastest time.

Gotcha. So, does it make sense to install them both, trying one for (how long, days, weeks?) and then uninstalling and installing the other? Do you think there would be a significant enough difference to go thru the trouble? If so, I guess it would probably make sense to try it before the card arrives, but if it arrives the 1st week or so of June, that is going to be SO hard to just let it lay there while the testing finishes.

---

Now here is the tricky part.

MB 8.12 also known as SoG is not part of Lunatics

So if you want to test SoG, you will need to edit the app_info.xml and app_config.xml and place those sections in before the cuda sections. Otherwise, it will default to cuda

Ok, that makes things a little easier. Now, is there a beta app that I should look to be using instead of the 'regular' stock app, which might be better for the new card? Or, just go with the current one?

---

Arg. Well, I can cross that bridge when I come to it when the card arrives.. :-)

---

Gotcha. So, does it make sense to install them both, trying one for (how long, days, weeks?) and then uninstalling and installing the other? Do you think there would be a significant enough difference to go thru the trouble? If so, I guess it would probably make sense to try it before the card arrives, but if it arrives the 1st week or so of June, that is going to be SO hard to just let it lay there while the testing finishes.

I would start with the AVX app. Then once you get the system fully built out with the GPUs you want and it is stable. Then give the SSE3 app a try.

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

Gotcha. So, does it make sense to install them both, trying one for (how long, days, weeks?) and then uninstalling and installing the other? Do you think there would be a significant enough difference to go thru the trouble? If so, I guess it would probably make sense to try it before the card arrives, but if it arrives the 1st week or so of June, that is going to be SO hard to just let it lay there while the testing finishes.

Thing is, we already know how the CPU run. So testing them is mostly for your own curiosity.

Can you install lunatics and run that before you put in the card and remove the lunatics when you want to install the card. Yes I believe you can.

Just make sure you use the uninstall that comes with it.

How long do you need to run each?

lol, some people are going to tell you need lots and lots of data, days to weeks.

I've looked at enough graphs in my time that I can start to see the patterns with a few hours where it is going. Including the outliers which people tend to forget always accompany any testing.

So I would tell you to do at least half a day to a full day if you aren't sure at each setting

The first numbers of actual crunching times of the 1080 final apears, at least on E@H.

https://einstein.phys.uwm.edu/hosts_user.php?userid=611401

Just not sure they are so impressive as expected.

Had a look at participant's posts to see if active in forums and stumbled on archae86's Pascal thread :)

For more info:
https://einstein.phys.uwm.edu/forum_thread.php?id=12101

BTW Al, you will love BoincTasks to see the change in run times :)

So much easier to read than the web page.

Brent, just got it installed on my main machine, it is definitely different, but I think I will like it once I've played around with it a little to get the look a little more to my liking. About all the remote machines, is there a small client piece I need to load onto each one, or do I do the full install like I did on this one, and they then automagically find each other, like one big happy family?

---

About all the remote machines, is there a small client piece I need to load onto each one, or do I do the full install like I did on this one, and they then automagically find each other, like one big happy family?

You don't need to install BoincTasks on any machine save the one on which you want to use it as a viewer. However you do need to tell the other machines to give permission for remote interrogation in the BoincTasks style.

There may be more than one way to do this. The way I have used for years is to create a text file named remote_hosts.cfg and place it in your BOINC directory (on the machine I am typing that directory is at c:\ProgramData\BOINC) but yours may be elsewhere). Each line of remote_hosts.cfg can contain a comment (by starting with a semicolon) or the ipv4 address of a PC from which you want to run BoincTasks, as viewable, for example by running ipconfig at a command line on a Windows machine.

One other thing, it seems to take restarting or rebooting on one or both ends of the connection before this permission takes effect.

Once you have established permission, you can add a computer to the set BoincTasks is displaying by any of an automatic scan, or using the network host name, or using the ipv4 address. In BoincTasks, select the Computer menu, then use either Add Computer or Find Computer.

The way I do things, I also need a password specification file at each remote PC, another text file, in the same directory, this one named gui_rpc_auth.cfg. The contents is a single line of text, constituting a password, which you also enter during the "add computer" dialog on the primary BoincTasks machine.

Again, other ways may work. This is one. You probably don't need all of this for the simpler monitoring functions, but I forget which is which. Fred's documentation is pretty good, so I suggest consulting it to straighten out any way I may have put you astray.

---

Thanks for the detailed response, only have one question from your reply. If I am running DHCP, not static IP addresses, and for whatever reason the IP addy on the viewer machine changes, will there be any way to automate the task of updating that info on all the other systems? Seeing as it is a text file, I'd guess no, but as you said, I'll dive into Freds documentation and check out the different options.

---

I beilve you can use computer name rather than IP, that would get around DHCP issues