This morning 5:00 hour the FB-DIMM sensor shows 64, roomtemp is 24 Celsius, nice temps for the V8-Xeon. ;-)

16 hours....It's hot today, roomtemp 28 Celsius, humidity 85%
FB-DIMM sensor 72 Celsius
CPU 0, Core 0 = 70, is the hottest
CPU 0, Core 3 = 63, is the coolest

I believe the differences come from the different parts of program are running.
Should nice to have a statistic about this problem.

heinz

16 hours....It's hot today, roomtemp 28 Celsius, humidity 85%
FB-DIMM sensor 72 Celsius
CPU 0, Core 0 = 70, is the hottest
CPU 0, Core 3 = 63, is the coolest

I believe the differences come from the different parts of program are running.
Should nice to have a statistic about this problem.

heinz

Which progs you let run.. not only SETI@home? ;-D
I have max. 10 - 12 °C difference between the 4 Cores at my QX6700. (100% SETI@home, 24/7)
Core0 is the hottest.

Difference between CPUs.. I have no experiences.. ;-D

---

Which progs you let run.. not only SETI@home? ;-D
I have max. 10 - 12 °C difference between the 4 Cores at my QX6700. (100% SETI@home, 24/7)
Core0 is the hottest.

Difference between CPUs.. I have no experiences.. ;-D

24/7 seti at all 8 cores under full load = 100%
Now after 10 days:
Pending credit: 11,249.15
Avg credit = 4,289.08
Total credit = 66,332

this looks like ~7500 credits/day

@Tsureku
This server is build to develop and test the HPF seti app. It is my working horse for all what I do ;-) if the burn in test are over.

heinz

Must now shut down all machines, a thunderstorm is comming

---

D5400XS V8-Xeon

HPF seti app? What's this?

Now the thunderstorm is away, time to switch ON again.. :-D

BTW.
Why you don't use some protection?
For example:
APC Performance SurgeArrest 8 outlets with Phone, Coax & Network Protection
or/and
APC SurgeArrest + Battery Backup

APC

---

If you use Crunch3rs BOINC V6.1.0 maybe you would have more performance..

He enabled CPU-affinity (every WU fixed to one Core) and RRI (report_results_immediately) (~ 60 seconds).

You have two seperate CPUs and in the CPUs two DUO-CPUs.. so normally you would see more crunching speed..


I use his client, because my QX6700 includes two DUO-CPUs.. ;-)
From my experiences the CPU is 1 - 2 °C cooler also..


Here at this board are much meanings about CPU-affinity enable or not..
Normally if you have two or more separate CPUs (one Core per CPU) you profit from it..
And from the architecture the Quad must profit from it also..

---

9:00 hour, we are back again, 86%humidity, 26 Celsius...looks like a hot day again.
FB-Dimm sensor shows 70 Celsius, cpus 49 to 65 Celsius..very different values
how we can explain it ?
@Sutaru
Thanks for the links to APC :-)
Sure a APC should be one of my next investments. I need a APC Smart-UPS 2200VA USB & Serial 230V (839,00 Euro) to be on the sure site.
I'm running Boinc V5.8.11 on my P4 2.66 Northwood, XP-home(32bit).
The V8-Xeon runs 64Bit OS, it runs BOINC V 5.10.45 without any problems.
HPF High Performance
HPC High Performance Computing(link on my website)
At the beginning of 2008 I compiled some test-HPF apps with ITBB, but had not the machine to test it...I asked in the forum at lunatics, but get no responce ;-(
If you want test it on your quaddy you can download from my website(at bottom).
Have a look at the Readme files before you start.
If you have questions, PM me.

heinz

....
At the beginning of 2008 I compiled some test-HPF apps with ITBB, but had not the machine to test it...I asked in the forum at lunatics, but get no responce ;-(
...

Neither did I have such a machine until beginning of february when my wolfdale upgrade came in, though I hope my Northwood times were Helpful ;D

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

....
At the beginning of 2008 I compiled some test-HPF apps with ITBB, but had not the machine to test it...I asked in the forum at lunatics, but get no responce ;-(
...

Neither did I have such a machine until beginning of february when my wolfdale upgrade came in, though I hope my Northwood times were Helpful ;D

Hi Jason,
sure your results are helpful...we two know the problematic :-)
It would be nice if you set up a small proxy for me in the next time, to go back to lunatics. (I believe the providers are not able or willing to solve the connection problems, they want make money to get through)
Or we must communicate over our "SVN TestProject".
At first I must reconfigure the V8-Xeon and install the developer environment. Think the next days I will get all I need todo so.
10 days full load at all 8 cores with 100%, the burn in test is over, the machine is very stable. (it was a good idee to choose the A+ case)
If the stress with the AKv8-porting is over, we can go back to our research task.

regards heinz :-)

Hi Jason,
sure your results are helpful...we two know the problematic :-)
It would be nice if you set up a small proxy for me in the next time, to go back to lunatics. (I believe the providers are not able or willing to solve the connection problems, they want make money to get through)
Or we must communicate over our "SVN TestProject".
At first I must reconfigure the V8-Xeon and install the developer environment. Think the next days I will get all I need todo so.
10 days full load at all 8 cores with 100%, the burn in test is over, the machine is very stable. (it was a good idee to choose the A+ case)
If the stress with the AKv8-porting is over, we can go back to our research task.

regards heinz :-)

Yep, hoping to get SSE out of the way this weekend, and then am looking at options & experiments for advancing SSE4.1 build
- incorporation of hand SSE4.1 in certain sections
- improved processor detection for improved affinity control reducing bus traffic (Internal)
- SSP helper threading to reduce cache misses (If Intel can solve a small data gathering problem there, they're working on it ;D)
- incorporation of some OpenMP in certain places, possible experiments moving toward a pipeline threading model that runs SaH stub apps driving an indepedant multithreaded pipeline instead of separate single threaded apps.
- some other ideas, which your notes have helped refine.

Jason

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

30.05.2008 07:13:27||Starting BOINC client version 5.10.45 for windows_x86_64
30.05.2008 07:13:27||log flags: task, file_xfer, sched_ops
30.05.2008 07:13:27||Libraries: libcurl/7.18.0 OpenSSL/0.9.8e zlib/1.2.3
30.05.2008 07:13:27||Executing as a daemon
30.05.2008 07:13:27||Data directory: C:\\Program Files\\BOINC
30.05.2008 07:13:27||BOINC is running as a service and as a non-system user.
30.05.2008 07:13:27||No application graphics will be available.
30.05.2008 07:13:27|SETI@home|Found app_info.xml; using anonymous platform
30.05.2008 07:13:28||Processor: 8 GenuineIntel Intel(R) Xeon(R) CPU E5405 @ 2.00GHz [Intel64 Family 6 Model 23 Stepping 6]
30.05.2008 07:13:28||Processor features: fpu tsc pae nx sse sse2 pni
30.05.2008 07:13:28||OS: Microsoft Windows Vista: , Service Pack 1, (06.00.6001.00)
30.05.2008 07:13:28||Memory: 8.00 GB physical, 16.04 GB virtual
30.05.2008 07:13:28||Disk: 50.00 GB total, 23.00 GB free
30.05.2008 07:13:28||Local time is UTC +2 hours
30.05.2008 07:13:28|SETI@home|URL: http://setiathome.berkeley.edu/; Computer ID: 4387433; location: home; project prefs: default
30.05.2008 07:13:28||General prefs: from SETI@home (last modified 20-May-2008 00:38:57)
30.05.2008 07:13:28||Host location: home
30.05.2008 07:13:28||General prefs: no separate prefs for home; using your defaults
30.05.2008 07:13:28||Preferences limit memory usage when active to 4094.35MB
30.05.2008 07:13:28||Preferences limit memory usage when idle to 7369.83MB
30.05.2008 07:13:28||Preferences limit disk usage to 23.00GB
30.05.2008 07:13:29||Running CPU benchmarks
30.05.2008 07:14:00||Benchmark results:
30.05.2008 07:14:00|| Number of CPUs: 8
30.05.2008 07:14:00|| 2024 floating point MIPS (Whetstone) per CPU
30.05.2008 07:14:00|| 5767 integer MIPS (Dhrystone) per CPU
30.05.2008 07:14:01|SETI@home|Restarting task 18ap08ah.18750.23385.13.8.103_0 using setiathome_enhanced version 528
30.05.2008 07:14:01|SETI@home|Restarting task 18ap08ah.18750.23385.13.8.25_0 using setiathome_enhanced version 528
30.05.2008 07:14:01|SETI@home|Restarting task 18ap08ah.18750.23385.13.8.28_0 using setiathome_enhanced version 528
30.05.2008 07:14:01|SETI@home|Restarting task 18ap08ah.18750.23385.13.8.49_0 using setiathome_enhanced version 528
30.05.2008 07:14:01|SETI@home|Restarting task 18ap08ah.18750.23385.13.8.59_0 using setiathome_enhanced version 528
30.05.2008 07:14:01|SETI@home|Restarting task 18ap08ah.18750.23385.13.8.99_0 using setiathome_enhanced version 528
30.05.2008 07:14:01|SETI@home|Restarting task 18ap08ah.18750.23385.13.8.6_0 using setiathome_enhanced version 528
30.05.2008 07:14:01|SETI@home|Restarting task 18ap08ah.18750.23385.13.8.88_0 using setiathome_enhanced version 528
30.05.2008 07:18:49|SETI@home|Computation for task 18ap08ah.18750.23385.13.8.103_0 finished
30.05.2008 07:18:49|SETI@home|Starting 18ap08ah.18750.23385.13.8.201_0
------------------------------------------------------------------------------------

9:00 hour, we are back again, 86%humidity, 26 Celsius...looks like a hot day again.
FB-Dimm sensor shows 70 Celsius, cpus 49 to 65 Celsius..very different values
how we can explain it ?

I know, many of our overclocked quaddies can do a wu in shorter times...and have better float point MIPS.
Keep in mind the E5405A is the smallest Xeon.
If the machine run stable as the last 10 days I will try to oc a bit to 3Ghz...if thermic and processor allow it.

the machine runs now 7500 credit/24h
can anybody of you post some credit/24h values ?
thanks

heinz

---

D5400XS V8-Xeon

Hello There Heinz

Here are the benchmarks for my Quaddie for comparison. It runs 32 bit XP with SP3

30/05/2008 13:56:37|SETI@home|Scheduler RPC succeeded [server version 601]
30/05/2008 13:56:37|SETI@home|Deferring communication for 11 sec
30/05/2008 13:56:37|SETI@home|Reason: requested by project
30/05/2008 13:58:17||Running CPU benchmarks
30/05/2008 13:58:18||Suspending computation - running CPU benchmarks
30/05/2008 13:58:48||Benchmark results:
30/05/2008 13:58:48|| Number of CPUs: 4
30/05/2008 13:58:48|| 4432 floating point MIPS (Whetstone) per CPU
30/05/2008 13:58:48|| 12438 integer MIPS (Dhrystone) per CPU
30/05/2008 13:58:50||Resuming computation


I am currently turning out 7,200 credits/24 hours on this rig, and the total RAC is still rising.

John

---

It's good to be back amongst friends and colleagues

Hi John,
your floatpoin MIPs are more than twice of my Xeon, impressed values, Congratulations.
So you can say you have a small Skulltrail.. ;-)

heinz

Hi John,
your floatpoin MIPs are more than twice of my Xeon, impressed values, Congratulations.
So you can say you have a small Skulltrail.. ;-)

heinz

heinz........you know it has been reported many times here that the reported benchmark scores are whacky........

But best o' luck with your SKT.....

---

"Time is simply the mechanism that keeps everything from happening all at once."

Hi John,
your floatpoin MIPs are more than twice of my Xeon, impressed values, Congratulations.
So you can say you have a small Skulltrail.. ;-)

heinz

As Mark says: - it might be an indication of potential, but the actual is where it is delivered. I am a litle disappointed the RAC is not higher yet despite running it for 3+ months, as I think the possibility is there (see other QX9650s).

---

It's good to be back amongst friends and colleagues

Hi John,
your floatpoin MIPs are more than twice of my Xeon, impressed values, Congratulations.
So you can say you have a small Skulltrail.. ;-)

heinz

As Mark says: - it might be an indication of potential, but the actual is where it is delivered. I am a litle disappointed the RAC is not higher yet despite running it for 3+ months, as I think the possibility is there (see other QX9650s).

The proof is in the crunching.........

The kitties never lie.........

---

"Time is simply the mechanism that keeps everything from happening all at once."

thanks to the kitties :-)

it's hot today, 28 Celsius roomtemp, 85% humidity, the FB-DIMM sensor shows 74 Celsius, the cpu' show 67 to 54 Celsius.

the next thunderstorm is comming, better I shut down now.

heinz

thanks to the kitties :-)

it's hot today, 28 Celsius roomtemp, 85% humidity, the FB-DIMM sensor shows 74 Celsius, the cpu' show 67 to 54 Celsius.

the next thunderstorm is comming, better I shut down now.

heinz

Crunchers do not like lightning.......... got some coming my way this afternoon, according to the forecast........

---

"Time is simply the mechanism that keeps everything from happening all at once."

18:42 we are back again,
the thunderstorm last night runs over 4 hours, such a lot of lightning I have never seen before, really a great "feu d`artifice". This one was a short, with a little bit rain...its now away over the border line direction Karlsruhe/Germany.
The problem is always if the lightning hit the 380KV power transmission line, then we lost all energy for a while , till the power-company swiched it on again. Therefore it is always better to shut down.
Hoping this one was the last for today.

It's now 27 Celsius and 88% humidity, FB-DIMM sensor shows 71 Celsius.
I'm still waiting for some tools, to reconfigure, but it looks like I don't get it this week.
So I let the V8-XEON crunch a little bit...

heinz