I don't see any doc for them when I tried to google them.

Thank you,
Tom

You've already read it. BOINC Client configuration

name
short name of the application as found in the corresponding <name>xxxxx</name> tags in your client_state.xml file.
The application name can also be found using the <sched_op_debug> logging flag: a new task shows "Starting task task_name using Application_Name ..."
max_concurrent
The maximum number of tasks of this application to run at a given time.
Each <app_version> element specifies parameters for a given app version; it overrides <app>. List-add.pngNew in 7.2.39
app_name
the short name of the application.
cmdline
command-line parameters to pass to the program.
project_max_concurrent
A limit on the number of running jobs for this project. List-add.pngNew in 7.4.9
report_results_immediately
If set, report this project's completed tasks immediately List-add.pngNew in 7.8.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

What file are max_project_concurrent and max_concurrent located in? I assume I have to find a "project." Under World Community Grid each one of the different sub-projects has its own project name. I am not sure how/where to put the above.

I don't see any doc for them when I tried to google them.

Thank you,
Tom

You would place those in app_config.xml in each project folder. Which is why I said, I don't think an "all_projects" would work in an app_config as those are limited to the specific project. In order for all projects to accept a command, they would think it need to be in the cc_config.xml. However, I've never tried such a command in the cc_config so I can't say for sure it would or would not work.

As for the max_concurrent and project _max_concurrent (sorry I reversed it in my original post)the exact position within the app_config.xml varies depending on which one you use.

project_max_concurrent is place in the second to last line of the app_config just before </app_config>

  
</app_version>
<project_max_concurrent>4</project_max_concurrent>
</app_config>

max_concurrent is place within the body of the app_confix just after the plan class example below

    
<app_name>setiathome_v8</app_name>
    <plan_class>opencl_nvidia_SoG</plan_class>
    <avg_ncpus>1</avg_ncpus>
    <ngpus>0.5</ngpus>
<max_concurrent>2</max_concurrent>

I can say that I haven't used the max_concurrent in a while as I use the project_max_concurrent. The only time I used to use the max_concurrent is when projects had more than 1 type of work unit, example Einstein or GPUGrid.

project_max_concurrent is place in the second to last line of the app_config just before </app_config>

  
</app_version>
<project_max_concurrent>4</project_max_concurrent>
</app_config>

Thank you.

---

A proud member of the OFA (Old Farts Association).

According to HP Support. The way HP Zxxx figures out if you have a "high performance" heat sink/fan is a jumper between pin 1 and pin 5

I have a stock Z400 heatsink/fan that has that jumper. I have just confirmed that the jumper is electrically active with an ohm meter.

I have previously installed the X5680 cpu and plugged in that cpu cooler.

The Z600 bio's complained about the cpu requiring "too much power" and halted there.

I hear by officially call the guess that an X5680 cpu should run on a stock Z600 motherboard with the latest bios and a "high performance" heatsink/fan to be wrong.

Ordering the X5675 cpus next.

Tom

---

A proud member of the OFA (Old Farts Association).

I hear by officially call the guess that an X5680 cpu should run on a stock Z600 motherboard with the latest bios and a "high performance" heatsink/fan to be wrong.

I should also note that this applies to a Z600 with a "boot block date" of 1/7/2010 which is the 2nd version of the motherboard. This mother board will run both 5500 and 5600 cpu's upto the X5675.

Tom

---

A proud member of the OFA (Old Farts Association).

Cpus, Memory and another cpu heatsink are all "due" on July 11th. I expect a busy Monday :)

Tom

---

A proud member of the OFA (Old Farts Association).

What file are max_project_concurrent and max_concurrent located in? I assume I have to find a "project." Under World Community Grid each one of the different sub-projects has its own project name. I am not sure how/where to put the above.

I don't see any doc for them when I tried to google them.

Thank you,
Tom

You would place those in app_config.xml in each project folder. Which is why I said, I don't think an "all_projects" would work in an app_config as those are limited to the specific project. In order for all projects to accept a command, they would think it need to be in the cc_config.xml. However, I've never tried such a command in the cc_config so I can't say for sure it would or would not work.

As for the max_concurrent and project _max_concurrent (sorry I reversed it in my original post)the exact position within the app_config.xml varies depending on which one you use.

project_max_concurrent is place in the second to last line of the app_config just before </app_config>

  
</app_version>
<project_max_concurrent>4</project_max_concurrent>
</app_config>

max_concurrent is place within the body of the app_confix just after the plan class example below

    
<app_name>setiathome_v8</app_name>
    <plan_class>opencl_nvidia_SoG</plan_class>
    <avg_ncpus>1</avg_ncpus>
    <ngpus>0.5</ngpus>
<max_concurrent>2</max_concurrent>

I can say that I haven't used the max_concurrent in a while as I use the project_max_concurrent. The only time I used to use the max_concurrent is when projects had more than 1 type of work unit, example Einstein or GPUGrid.

Using project_max_concurrent is what I initially suggested they use, but figured they had their reasons for specifying each app.
I guess it wasn't helpful that everyone was listening the wrong value of max_project_concurrent immediately afterward...

I'm curious as to the "all_projects" value you keep keep mentioning. Is meant to refer to the all_apps value I said I would like if it existed?
I do think it would be nice to be able to define some other parameters, such as GPU/CPU usage, for all apps in a project at once.

---

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

I'm curious as to the "all_projects" value you keep keep mentioning. Is meant to refer to the all_apps value I said I would like if it existed?

I'm thinking he wanted to limit the total from all the different projects that would be running at any time on his computer. Such a command doesn't exist if I understand this correct. I could be wrong. If such a command existed, I would think it would have to be higher up in the folders than the project folder. That is why I speculated that it would probably be needed to be in the cc_config.xml as all projects read that file. But as to how it would determine which projects ran how many??? I don't know. I think the project_max_concurrent makes more sense ( as you suggested) as you can tailor it to the number for each project and from there figure out which you want to run.

I do think it would be nice to be able to define some other parameters, such as GPU/CPU usage, for all apps in a project at once.

This statement confused me as you already can define preferences for GPU to CPU ratio in the app_config.xml. Note, I said preference as system will use what it needs based on the application. So it may use more if it determines that is what is needed. Were you meaning something else???

I'm curious as to the "all_projects" value you keep keep mentioning. Is meant to refer to the all_apps value I said I would like if it existed?

I'm thinking he wanted to limit the total from all the different projects that would be running at any time on his computer. Such a command doesn't exist if I understand this correct. I could be wrong. If such a command existed, I would think it would have to be higher up in the folders than the project folder. That is why I speculated that it would probably be needed to be in the cc_config.xml as all projects read that file. But as to how it would determine which projects ran how many??? I don't know. I think the project_max_concurrent makes more sense ( as you suggested) as you can tailor it to the number for each project and from there figure out which you want to run.

I do think it would be nice to be able to define some other parameters, such as GPU/CPU usage, for all apps in a project at once.

This statement confused me as you already can define preferences for GPU to CPU ratio in the app_config.xml. Note, I said preference as system will use what it needs based on the application. So it may use more if it determines that is what is needed. Were you meaning something else???

Other than changing the CPU limit settings. I don't think BOINC has a way to limit the total number of tasks running at once.
It can get a bit more complicated if you only ever want BOINC to use 4 out of 8 CPUs for CPU processing and the rest reserved for GPUs even if there are 0 GPU tasks on the system.
In that scenario I set ncpus to 4 in the cc_config and then set cpu_usage to 0.001 in the app_config for gpu_versions. Logically that should work but sometimes BOINC has its own ideas.


For the GPU/CPU usage. If you want to add usage parameters it is per application.
Like this:

<app_config>
	<app>
		<name>setiathome_v8</name>
		<gpu_versions>
			<gpu_usage>0.5</gpu_usage>
			<cpu_usage>0.5</cpu_usage>
		</gpu_versions>
	</app>
	<app>
		<name>setiathome_v9</name>
		<gpu_versions>
			<gpu_usage>0.5</gpu_usage>
			<cpu_usage>0.5</cpu_usage>
		</gpu_versions>
	</app>
<app_config>

As some projects have a large number of apps. I was thinking that having a way to define parameters for app apps in a project would be useful.
Where some predefined "name" could be used if one setting for all apps in a project was desired.
Something along the lines of:

<app_config>
	<app>
		<name>all_apps</name>
		<gpu_versions>
			<gpu_usage>0.5</gpu_usage>
			<cpu_usage>0.5</cpu_usage>
		</gpu_versions>
	</app>
<app_config>

Maybe something like a project_config.xml is really what would be needed...

---

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

I am pleased to announce that I have just put 24 "cores" (12 cores w/HT) online for "about" $505.99

Z600 w/E5620 cpu, 6 GB Ram, no HD, no OS $164.00
2 X5675 cpus (3.06Ghz, 6 cores w/HT) 100.00
2 1 terabyte HD (Raid1) [transferred in] ~100.00
1 Z600 z800 cpu cooler 41.99
1 Gtx 750 TI (for pricing purposes, ~100.00
actually have 2 Gtx 1060's installed ========
~$505.99

I do lust after 32 cores but this was within my budget.

Tom

---

A proud member of the OFA (Old Farts Association).

It is being "enforced" at server level, using data provided from the client.

Yup,
1* And the client can tell what is "has" and it can issue the work to what it really has..
2* Then there is another limit in the client. That is 1000 WU's. Search for 1000 not 1024 in the boinc client source code and compile your own.

With a high CPU core count the 1000 WU limit can become an issue too.

Luckily we all run on public source code environment projectwise.

I did my mod because of the Tuesdays. Hence the 16 * 1080 Tu. Three 1080 and one 1080Ti really. Just to survive the outage.

P.

---

To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones

I just got done "pricing" the 32-Thread Xeon and it looks like you could get this machine (less the GPU's) for around $1,363 plus s/h.

https://www.techspot.com/review/1155-affordable-dual-xeon-pc/

Memory G.SKILL Ripjaws Z Series 64GB (8 x 8GB) 240-Pin DDR3 SDRAM DDR3 1866 (PC3 14900) $473~ new egg

ASRock EP2C602-2T/D16 SSI EEB Server Motherboard Dual LGA 2011 DDR3 1600/1333/10 $295~ on ebay

Intel Xeon E5-2670 2.6GHz Eight Core (BX80621E52670) Processor $99~ x 2 on ebay

cpu coolers Noctua NH-U12DX i4 $64~ X 2 on newegg

CORSAIR RMx Series RM1000X 1000W 80 $169~

Case about $100

Total about: $1363

Assuming consumer ram.

Oh well, I've already blown my birthday budget for this year.

Tom

---

A proud member of the OFA (Old Farts Association).

Sigh, I was mucking around and found a turnkey just an OS.

https://www.newegg.com/Product/Product.aspx?Item=9SIAC0F4WC9675&ignorebbr=1&nm_mc=KNC-GoogleMKP-PC&cm_mmc=KNC-GoogleMKP-PC-_-pla-_-Server%2FWorkstation+Systems-_-9SIAC0F4WC9675&gclid=CjwKCAjw16HLBRBFEiwAElREqDgrAaMd4thaljRfFhoQMBPDJnugfMeTpIzaLlr3Mk3nsCCRFPdoKxoCa_UQAvD_BwE&gclsrc=aw.ds

It even has 128GB of ram.

Tom

---

A proud member of the OFA (Old Farts Association).

I just got done "pricing" the 32-Thread Xeon and it looks like you could get this machine (less the GPU's) for around $1,363 plus s/h.

https://www.techspot.com/review/1155-affordable-dual-xeon-pc/

Memory G.SKILL Ripjaws Z Series 64GB (8 x 8GB) 240-Pin DDR3 SDRAM DDR3 1866 (PC3 14900) $473~ new egg

ASRock EP2C602-2T/D16 SSI EEB Server Motherboard Dual LGA 2011 DDR3 1600/1333/10 $295~ on ebay

Intel Xeon E5-2670 2.6GHz Eight Core (BX80621E52670) Processor $99~ x 2 on ebay

cpu coolers Noctua NH-U12DX i4 $64~ X 2 on newegg

CORSAIR RMx Series RM1000X 1000W 80 $169~

Case about $100

Total about: $1363

Assuming consumer ram.

Oh well, I've already blown my birthday budget for this year.

Tom

You might have missed the link to this site in the other thread about building a 32 threaded system.
http://natex.us/intel-s2600cp2j-motherboard-dual-e5-2670-sr0kx-64gb-pc3-12800r/
Pretty much the same specs but about $436 less. Last October/November I was buying the version bundled with 128GB of ram and they were only like $480 each instead of $740. Supply and demand at work!

However this looks like a better packaged deal. Since it is everything except HDDs for $485 less and 128GB instead of 64GB of ram.
http://natex.us/enthoo-128kit

They also have rack servers if that is your thing.
http://natex.us/servers/servers-package-deals/

---

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

However this looks like a better packaged deal. Since it is everything except HDDs for $485 less and 128GB instead of 64GB of ram.
http://natex.us/enthoo-128kit

Pant, Pant, Pant looks lustfully at 36 more cores.... :)

I like at the package deal because I do have spare HD's and can always reallocate a Gtx 1060. It lists it as "out of stock" so I wonder how long a backorder would be? Hmmmmm......

That really looks like it could be within my price range... :)

Tom

---

A proud member of the OFA (Old Farts Association).

You might have missed the link to this site in the other thread about building a 32 threaded system.
http://natex.us/intel-s2600cp2j-motherboard-dual-e5-2670-sr0kx-64gb-pc3-12800r/
Pretty much the same specs but about $436 less. Last October/November I was buying the version bundled with 128GB of ram and they were only like $480 each instead of $740. Supply and demand at work!

However this looks like a better packaged deal. Since it is everything except HDDs for $485 less and 128GB instead of 64GB of ram.
http://natex.us/enthoo-128kit

They also have rack servers if that is your thing.
http://natex.us/servers/servers-package-deals/

I did miss that link. I have since visited it and the price has jumped just a little bit to $878.00 While it is currently "out of stock" I just got an email back from them that said they will get fans tomorrow and have it in stock again.

My main concern about the video card slots is the "!Please note!
This does not have a 16x slot".

Since this would be for a General Purpose Boinc cruncher but would be urged on with at least a Gtx 1060 video card, I am worried about the limited number of high high end video slots. And wondering if anyone knows if the "not having a 16x slot" would slow down Seti gpu processing? I have heard of one gpu intensive project where the bandwidth between the cpu and the gpu does make a difference (Milky way or maybe Einstein? Don't remember since it wasn't a project I am doing).

Thank you,
Tom

---

A proud member of the OFA (Old Farts Association).

I have also been poking around e-bay and when searching for an "dual cpu e5-2670 workstation" have run across some Z800's? configured to be turnkey (only 32 GB ram?) some even have hard disks.

I think I remember some below $1,000 although that was a very common price. These usually have 1 TeraB HD's and other sometimes useful peripherals and almost none have a video card.

Here is one to lust after: HP Z620 Workstation 16-Core 2.60GHz E5-2670 32GB RAM 2x 500GB HDD Buy it now for $US $769.99 with free s/h.

Tom

---

A proud member of the OFA (Old Farts Association).

You might have missed the link to this site in the other thread about building a 32 threaded system.
http://natex.us/intel-s2600cp2j-motherboard-dual-e5-2670-sr0kx-64gb-pc3-12800r/
Pretty much the same specs but about $436 less. Last October/November I was buying the version bundled with 128GB of ram and they were only like $480 each instead of $740. Supply and demand at work!

However this looks like a better packaged deal. Since it is everything except HDDs for $485 less and 128GB instead of 64GB of ram.
http://natex.us/enthoo-128kit

They also have rack servers if that is your thing.
http://natex.us/servers/servers-package-deals/

I did miss that link. I have since visited it and the price has jumped just a little bit to $878.00 While it is currently "out of stock" I just got an email back from them that said they will get fans tomorrow and have it in stock again.

My main concern about the video card slots is the "!Please note!
This does not have a 16x slot".

Since this would be for a General Purpose Boinc cruncher but would be urged on with at least a Gtx 1060 video card, I am worried about the limited number of high high end video slots. And wondering if anyone knows if the "not having a 16x slot" would slow down Seti gpu processing? I have heard of one gpu intensive project where the bandwidth between the cpu and the gpu does make a difference (Milky way or maybe Einstein? Don't remember since it wasn't a project I am doing).

Thank you,
Tom

All of the PCie slots on the Intel S2600CP2J are only x8 electrical. Given in most dual GPU setups they normally end up running at x8 anyway I don't see it as a huge deal.
My original plan was to start with a pair of 750 Ti's or 1050 Ti's and then use PCIe risers/extenders to run 2 more. However a pair of 1060 3GB cards was looking more cost effective. At least before this shortage nonsense started...

There are some projects that do tend to use more PCIe bandwidth than others. I believe SETI@home is on the lower usage end of the scale. My 750 Ti is currently running in a PCIe x1 slot that is only v2. Which hasn't effected its performance vs running in a x16 slot. A larger GPU like a 1080 ti may see some change.
At the moment I'm running Einstein tasks on my R9 390x and the PCIe utilization is fairly low. I run Milkyway as well and will have to double check it's usage once I finish the Einstein tasks. I also don't know hoe comparable the PCIe usage is between Radeon and Nvidia GPUs.

---

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

However this looks like a better packaged deal. Since it is everything except HDDs for $485 less and 128GB instead of 64GB of ram.
http://natex.us/enthoo-128kit

........
I did miss that link. I have since visited it and the price has jumped just a "little bit" to $878.00 While it is currently "out of stock" I just got an email back from them that said they will get fans tomorrow and have it in stock again.

I (gulp) bit the bullet and ordered a "purpose built" machine for the first time. The cpus on this one are late model enough to have AVX extensions so it is MUCH faster on its cpu processing. Its due about next tuesday. I will need to move a HD and probably my Gtx 1060 into it (I could start with a Gtx 750Ti but why delay the inevitable?). As well as shutdown the other two pc's to get my AC bill back into some kind of semblance of order.

Tom

---

A proud member of the OFA (Old Farts Association).

http://natex.us/enthoo-128kit

I (gulp) bit the bullet and ordered a "purpose built" machine for the first time. The cpus on this one are late model enough to have AVX extensions so it is MUCH faster on its cpu processing. Its due about next Tuesday. I will need to move a HDD and probably my Gtx 1060 into it (I could start with a Gtx 750Ti but why delay the inevitable?). As well as shutdown the other two pc's to get my AC bill back into some kind of semblance of order.

Tom

Well I now have the above machine up and computing.

Notes on getting it setup.
1) Those little "beeps" that show up every time you boot it are the glad cries of the MB recognizing that the USB ports are up and running.
2) If/when you put in a discrete video card. Please install it in the furthest away from the cpu blue slot instead of spending a half hour trying to figure out why your known working video card wasn't working while installed in the blue slot closest to the cpus. I even used a 2nd known working video card to get the same non-working results before it occurred to me to try the other blue slot.
3) Don't try to migrate a HDD from 1/2 of a raid 1 system so you don't have to install from scratch. In my case it also produced a mysterious rattle noise that I still don't see how it did it. Installing Win7 from scratch on the same HDD after deleting all the partitions made things MUCH cleaner.
4) Expect you will need motherboard specific cpu chip drivers as well as Network/Lan drivers. You will need to download those from Intel on another machine. The MB is a "S2600CP" or there abouts.
5) Don't download a Nividia driver package for Win10 and wonder why it is "incompatible" with your Win7 install.
6) If you aren't going to do gpu processing under Boinc you don't need a video card. The MB does have one built.
7) Unless you have a very tall monitor or want to run it in very high resolution, your task listing under Boinc will scroll off the bottom of your screen :D

HTH,
Tom

---

A proud member of the OFA (Old Farts Association).