Posts by fractal
log in
1) Message boards : Number crunching : Longer Sieve WUs (Message 23597)
Posted 398 days ago by fractal
I do NOT want to go back to having multiple sizes of WUs since most people never selected the appropriate size for their machine which caused lots of complaints about not getting the ones they wanted.

If we were voting, I would vote for a optional larger WU with appropriately larger credit. Donors would see a credit from reducing the number of turnarounds. No need to increase it.

And, even though you say you do NOT want to go to it, I would make the default the current set of work. Donors MUST select the newer units if they want them. And, those who can not figure out how to select them, or can not figure out how to run optimized settings can go bleep themselves ;)

So, default doesn't change. AFK drivers continue to get what they got. Attentive donors can get longer work some of the time. The only cost is an additional work queue.
2) Message boards : Number crunching : Micro C. for Android/x86 (Message 23588)
Posted 399 days ago by fractal
I thought all X86 code was position independent. Probably just a matter of removing the flag.

I put an x86_64 android VM on this project. It is not getting any work. It will accept either x86 or x86_64 work since it is requesting:


Boinc is pretty clear about PIE at
3) Message boards : News : SSL Certificate Replaced (Message 23302)
Posted 437 days ago by fractal

The following script let my Ubuntu Linux machine fetch work.

cd /etc/ssl/certs/
sudo wget
sudo ln -s sca.server1.crt 7f8496de.0

It downloads the Startcom intermediate certificate and installs it in the ssl certificate library.

Startcom has documentation on how the web server can publish the intermediate certificate to avoid this if you have access to the web server. explains it in detail and explains why checking in your browser does not help boinc which uses libcurl.

Sorry, but windows users are on their own to figure out where to put the certificate.
4) Message boards : News : SSL Certificate Replaced (Message 23300)
Posted 437 days ago by fractal
This doesn't look like a DNS issue. It looks like a certificate issue.

I got it to work temporarily by editing client_state.xml and replacing https with http for collatz. Unfortunately it reset itself overnight and I am back to getting no work.

Boinc uses libcurl to access the site so it is very easy to debug

curl -v
* Hostname was NOT found in DNS cache
* Trying
* Connected to ( port 443 (#0)
* successfully set certificate verify locations:
* CAfile: none
CApath: /etc/ssl/certs
* SSLv3, TLS handshake, Client hello (1):
* SSLv3, TLS alert, Client hello (1):
* SSLv3, TLS handshake, Server hello (2):
* SSLv3, TLS handshake, CERT (11):
* SSLv3, TLS alert, Server hello (2):
* SSL certificate problem: unable to get local issuer certificate
* Closing connection 0
curl: (60) SSL certificate problem: unable to get local issuer certificate
More details here:

curl performs SSL certificate verification by default, using a "bundle"
of Certificate Authority (CA) public keys (CA certs). If the default
bundle file isn't adequate, you can specify an alternate file
using the --cacert option.
If this HTTPS server uses a certificate signed by a CA represented in
the bundle, the certificate verification probably failed due to a
problem with the certificate (it might be expired, or the name might
not match the domain name in the URL).
If you'd like to turn off curl's verification of the certificate, use
the -k (or --insecure) option.

So, the issuer's CA isn't trusted. Let's see what the checkers think. thinks your trust chain isn't very good, just like curl.

Neither does

You may be able to convince your volunteers to install the StartCom Class 1 certificate in their systems so they can connect, but it may be better if you get a certificate from a trusted authority.

Google is your friend btw. See
5) Message boards : Unix/Linux : LLVM ERROR? (Message 21971)
Posted 721 days ago by fractal
It has been a couple of years since this was reported. Has anyone figured it out yet? I tried adding one of my ATI machines to the project and am seeing

Stderr output

process exited with code 1 (0x1, -255)
Collatz Conjecture Sieve 1.21 Linux x86_64 for OpenCL
Written by Slicker (Jon Sonntag) of team SETI.USA
Based on the AMD Brook+ kernels by Gipsel of team Planet 3DNow!
Sieve code and OpenCL optimization provided by Sosiris of team BOINC@Taiwan
LLVM ERROR: Cannot select: 0x1fe72c0: i32 = setcc 0x22836f0, 0x2281ce0, 0x1fe71c0 [ORD=14] [ID=50]
0x22836f0: i64 = add 0x2281ce0, 0x22829e0 [ORD=10] [ID=46]
0x2281ce0: i64 = AMDILISD::VEXTRACT 0x23a8ee0, 0x23a85e0 [ORD=9] [ID=36]
0x23a8ee0: v2i64,ch = CopyFromReg 0x267bfb0, 0x23a8de0 [ORD=9] [ID=30]
0x23a8de0: v2i64 = Register %vreg38 [ORD=9] [ID=2]
0x23a85e0: i32 = TargetConstant<1> [ORD=2] [ID=21]
0x22829e0: i64 = add 0x22826e0, 0x22828e0 [ORD=8] [ID=45]
0x22826e0: i64,ch = load 0x267bfb0, 0x22821e0, 0x22822e0<LD4[%arrayidx], zext from i32> [ORD=5] [ID=44]
0x22821e0: i32 = add 0x23a88e0, 0x22820e0 [ORD=3] [ID=43]
0x23a88e0: i32,ch = CopyFromReg 0x267bfb0, 0x23a87e0 [ORD=3] [ID=29]
0x23a87e0: i32 = Register %vreg35 [ORD=3] [ID=1]
0x22820e0: i32 = shl 0x1fe6cc0, 0x2281fe0 [ORD=3] [ID=42]
0x1fe6cc0: i32 = AMDILISD::VEXTRACT 0x23a92e0, 0x23a85e0 [ORD=2] [ID=38]
0x23a92e0: v4i32 = 0x23a91e0 [ORD=1] [ID=32]
0x23a91e0: i32 = TargetConstant<2975> [ORD=1] [ID=4]
0x23a85e0: i32 = TargetConstant<1> [ORD=2] [ID=21]
0x2281fe0: i32 = Constant<2> [ORD=3] [ID=7]
0x22822e0: i32 = undef [ORD=4] [ID=8]
0x22828e0: i64 = shl 0x22825e0, 0x22827e0 [ORD=7] [ID=40]
0x22825e0: i64 = zero_extend 0x23a90e0 [ORD=6] [ID=37]
0x23a90e0: i32,ch = CopyFromReg 0x267bfb0, 0x23a8fe0 [ORD=6] [ID=31]
0x23a8fe0: i32 = Register %vreg39 [ORD=6] [ID=3]
0x22827e0: i32 = Constant<26> [ORD=7] [ID=9]
0x2281ce0: i64 = AMDILISD::VEXTRACT 0x23a8ee0, 0x23a85e0 [ORD=9] [ID=36]
0x23a8ee0: v2i64,ch = CopyFromReg 0x267bfb0, 0x23a8de0 [ORD=9] [ID=30]
0x23a8de0: v2i64 = Register %vreg38 [ORD=9] [ID=2]
0x23a85e0: i32 = TargetConstant<1> [ORD=2] [ID=21]
In function: __OpenCL_kernelSteps_kernel


Which look remarkably like what the OP reported. Nothing interesting is reported by ldd
:~/BOINC/projects/boinc.thesonntags.com_collatz$ ldd -v collatz_sieve_1.21_x86_64-pc-linux-gnu__opencl_amd_gpu => (0x00007fff28bfe000) => /lib/x86_64-linux-gnu/ (0x00007fcba90cd000) => /usr/lib/x86_64-linux-gnu/ (0x00007fcba8ec7000) => /lib/x86_64-linux-gnu/ (0x00007fcba8bca000) => /lib/x86_64-linux-gnu/ (0x00007fcba880c000)
/lib64/ (0x00007fcba9309000) => /lib/x86_64-linux-gnu/ (0x00007fcba8608000)

Version information:
./collatz_sieve_1.21_x86_64-pc-linux-gnu__opencl_amd_gpu: (GLIBC_2.3) => /lib64/ (GLIBC_2.2.5) => /lib/x86_64-linux-gnu/ (OPENCL_1.0) => /usr/lib/x86_64-linux-gnu/ (GLIBC_2.2.5) => /lib/x86_64-linux-gnu/ (GLIBC_2.4) => /lib/x86_64-linux-gnu/ (GLIBC_2.3) => /lib/x86_64-linux-gnu/ (GLIBC_2.2.5) => /lib/x86_64-linux-gnu/
/lib/x86_64-linux-gnu/ (GLIBC_2.2.5) => /lib64/ (GLIBC_2.3) => /lib64/ (GLIBC_PRIVATE) => /lib64/ (GLIBC_2.14) => /lib/x86_64-linux-gnu/ (GLIBC_2.3.2) => /lib/x86_64-linux-gnu/ (GLIBC_PRIVATE) => /lib/x86_64-linux-gnu/ (GLIBC_2.2.5) => /lib/x86_64-linux-gnu/
/usr/lib/x86_64-linux-gnu/ (GLIBC_2.2.5) => /lib/x86_64-linux-gnu/ (GLIBC_2.2.5) => /lib/x86_64-linux-gnu/
/lib/x86_64-linux-gnu/ (GLIBC_PRIVATE) => /lib/x86_64-linux-gnu/ (GLIBC_2.2.5) => /lib/x86_64-linux-gnu/
/lib/x86_64-linux-gnu/ (GLIBC_2.3) => /lib64/ (GLIBC_PRIVATE) => /lib64/
/lib/x86_64-linux-gnu/ (GLIBC_PRIVATE) => /lib64/ (GLIBC_PRIVATE) => /lib/x86_64-linux-gnu/ (GLIBC_2.2.5) => /lib/x86_64-linux-gnu/

I have had this machine on moo for a while but it is down today so tried to add collatz. I had a similar failure when I tried many months ago. I have it crunching Einstein@home work now.
6) Message boards : Number crunching : Optimized Applications (Message 6465)
Posted 2875 days ago by fractal is where you download the application. Instructions are in the readme file.
7) Message boards : Number crunching : PSU's?? (Message 6421)
Posted 2878 days ago by fractal
If your building a BOINC only Box you can get away with a lot less PSU than if your were building a Gaming Box.

Great point! Someone building a gaming box with plans on adding a second gpu, or perhaps upgrading in a year might very well go a size or two up on the PSU. But, this comes at a cost.

An 80+ rated supply is required to be efficient all the way down to 20% load. The OP's system, while it probably needs just over 400w when crunching, likely consumes close to 100w when idle (back of the napkin guestimates). That does not leave a whole lot of wiggle room if you want to stay in the 80% efficient range both idle and loaded.

A crunch box doesn't have these concerns. They are rarely idle, so efficiency at idle is of little concern. A crunch box wants to aim for 60-70% load. This allows the PSU fan to stay reasonably quiet (in most benchmarks) without costing too very much more. But someone's who games but does not crunch, who's box is idle most of the time, might just worry about the extra 10-20 watts lost to heat when idle by oversizing the PSU.
8) Message boards : Number crunching : PSU's?? (Message 6399)
Posted 2880 days ago by fractal
My PC has 2 HD5850 and a i7-920 and 'only' a 750W PSU and everthing is working fine. There was only one problem: There are only few PSUs which have 4 6-pin connectors!

Included in one of my boxes with a new video card was a 2 into one connector, it took 2 molex plugs, the ide drive power plugs, and combined them into one 6 pin plug for the video card. That means you should be able to find them online.

A couple of things. First of all, buying a psu with 4 x 6-pin PCIe power connectors is the "safe" way to go. This suggests the manufacturer intended it to run a pair of high power cards. Of course you can use the adapters and some times it will even work.

Next, reading the numerous posts like this is source of great amusement. People who think you need 1000 watt power supplies to run a system like the op which will barely break 400 watts running flat out.

So ... safe advice remains to buy a power supply from a reputable company with enough PCIe power connectors to do the job. You will want to make sure you get an 80plus rated supply for multiple gpu applications. Chances are that the engineer who designed the power supply built it with the capacity to drive the number of plugs. This will usually give you a unit operating in the sweet 40-60% load range. Given your system will draw around 400 watts, aiming for 60% load suggest you will want a properly designed 650 watt unit. 850 is not that much overkill and most 850 watt supplies come with 4 PCIe power plugs. 1000 is just a waste. The 90 dollar pc power and cooling silencer 750 is an excellent choice for the system you describe. Keep that in mind if what you find costs more.

Alternately, you can use the molex to pcie adapters. You can either know how many rails you have, where they are presented and what their capacity is, or you can hope you get lucky. If your supply has two PCIe plugs on one cable, plugging one of those plugs into each card and making up the other two with adapters has a better chance of succeeding than plugging both into one card and feeding the other card with two adapters. Knowing how your rails are presented is, of course, even better.

I encourage anyone who likes to play with systems to pick up a kill-a-watt device. They cost about 20 dollars. These will help you make informed decisions in terms of how big a power supply you need for your box, whether you should perhaps buy a smaller supply for your current box and use that 1000 watt supply you bought based on forum advice for your new quad GPU system, and how many systems you can safely run off your wall plug.

Finally, for my specifics. I have two dual gpu boxes. They use more modest GPU's than the OP, one a pair of 8800gt's and the other a pair of 9600gso's. Both systems have low power x2 processors. Both systems have antec ea-380 PSU's. Both use the molex->pcie power plugs that came in the box with the cards. Both have been crunching for more than a year. Both draw around 200 watts from the wall when crunching.
9) Message boards : Number crunching : Completed, validation inconclusive (Message 62)
Posted 3115 days ago by fractal
That was a hint, perhaps too subtle, that slicker might want to detect invalid results...
10) Message boards : Number crunching : Completed, validation inconclusive (Message 60)
Posted 3115 days ago by fractal
I switched to 185.85 and it completed a work unit in about 9 minutes, more typical of what I have seen before. The downside is this machine is a LOT more slugish, particularly when gaming.

What is more interesting is that one of my work units got a similarly broken wingman, and as such, we both got credit for it.
11) Message boards : Number crunching : Completed, validation inconclusive (Message 57)
Posted 3115 days ago by fractal
Did you mean 185.85? The only .50 I see is 182.50. There is some debate on the gpugrid forums whether 185.85 is good or not. I still have the old install files for the ancient driver I am running, so I think I will give 185.85 a try.
12) Message boards : Number crunching : Completed, validation inconclusive (Message 53)
Posted 3115 days ago by fractal
Yes, I need to upgrade both boinc and the drivers on that machine. I settled on that combination many months ago as it was stable and worked with GPUGRID. Finding the perfect match of forceware and boinc without having to tweak it daily is difficult. It took me several hours of reading the forums to find the combination that had the least number of complaints against it.
13) Message boards : Number crunching : Completed, validation inconclusive (Message 50)
Posted 3115 days ago by fractal
ForceWare 177.35

aqua@home didn't like that machine either. Guess I will leave it on gpugrid until I have time to do major maintenance on it.
14) Message boards : Number crunching : Completed, validation inconclusive (Message 47)
Posted 3115 days ago by fractal
I am getting a bunch of these. is an example. Oddly, my machine finished the unit in 28 seconds, the other machine, a slightly more powerful card, took 750 seconds.

Main page · Your account · Message boards

Copyright © 2018 Jon Sonntag; All rights reserved.