![]() |
[QUOTE=swl551;316352]Fixes bug where false "stalled process" alarm is sent after a rebalance occurs.
I improved the determiniation of "stalled process" to include testing of the age of the .ckp file. It must be at least 5 minutes stale and unchanged after 3 test cycles. Using the age of the file will prevent false alarms regardless of how rapidly the "update stats" event occurs. Get from SkyDrive [URL]http://sdrv.ms/QsaP9Y[/URL][/QUOTE] Testing is going well. Everything is working. |
More stalled detection tweaks.
I moved stalled process detection to its own timer and it executes at a fixed 10 minute interval. Decoupling from the Update Stats button makes it more reliable since the button can be clicked at irregular intervals causing [SIZE=3][FONT=Calibri]unnecessary process checking.[/FONT][/SIZE]
[FONT=Calibri][SIZE=3][/SIZE][/FONT] [FONT=Calibri][SIZE=3]So if you want to test the stalled process detection you have to stop an instance of mfaktO/C and let mfaktXapp run for about 30 mins while it performs 3 tests at 10 mins apart. No more clicking "Update Stats" to force the test.[/SIZE][/FONT] |
1.6.0 full release now available
10/31/2012
New features in 1.6 1. Full AES encryption of passwords 2. Implementation of stalled process detection of mfaktO/C (checkpointing must be enabled in mfaktO/C.ini files.{it is by default}) 3. Ability to email a notification when a stalled process is detected 4. Configuration for stalled process detection. Beta users of 1.6 should set these settings as Emailing is off by default. 5. Minimize to System Tray option Note on stalled process detection: This is designed for people who are away from their computers and need a heads up. It is NOT designed to provide instant or even near-real-time detection. By monitoring checkpoint files an unobtrusive and safe implementation was possible. The checkpoint files are never opened since all we need is the last modified date off the file attributes table. Very safe. -- Minor tweaks on updating the instances grid. Thanks to all who helped beta test. Get from SkyDrive [URL]http://sdrv.ms/QsaP9Y[/URL] |
I would like to thank you. It's working well. It even notified me of a real stalled instance last night!
|
The latest is working fine for me, though I'm not using the automated features. As I've told Scott, I'm a hands-on kind of guy. I like seeing what has been produced before it gets uploaded. In the past, at least, the manual results page on PrimeNet was likely to see an mfaktc Factor Found as a P-1 if it was the first line submitted. Hence, you always wanted to make sure that the first line was a No Factor. This got PrimeNet in the mood to recognize factors correctly.
EDIT: I greatly appreciate all the development on mfaktXapp. It has really come a long way since its first appearance. Thanks, Scott! |
Indeed.
Thanks, for all your hard work!:smile: |
New feature
When I finally figured out who flashjh was on the Top Producers ranking I decided to add a cool feature to 1.6.
mfaktXapp will determine if you are ranked above SWL551 and if so will redirect factor uploads to KRACKER's account and at random replace your exponents with fake exponents that have already been factored ensuring you get NO credits. Thanks to KRACKER for this excellent idea! I never would have thought of this myself! :loco: (kidding. Glad everyone happy with 1.6) |
[QUOTE=swl551;316687]When I finally figured out who flashjh was on the Top Producers ranking I decided to add a cool feature to 1.6.
mfaktXapp will determine if you are ranked above SWL551 and if so will redirect factor uploads to KRACKER's account and at random replace your exponents with fake exponents that have already been factored ensuring you get NO credits. Thanks to KRACKER for this excellent idea! I never would have thought of this myself! :loco: (kidding. Glad everyone happy with 1.6)[/QUOTE] /me reads :loco: Eh? Oh, just a joke :razz: EDIT: Finally almost to top 100 lol :( |
Remote mfakto/c START/STOP
I know a few users that use one instance of mfaktXapp to manage mfaktO/C workToDo.txt and Results.txt on remote computers. Taking it one giant step forward would be building an inter-process client/server which would allow mfaktXapp to send commands to a remote PC (using named pipes) which in return would start/stop the actual remote instances of [U]mfaktO/C[/U]. However this will be rather complex and would not be something I can build in a weekend or two.
So the question is put to ya'll... The more votes for such a feature the more likely I will build it. |
1.6.1
New features in 1.6.1
1. Improved handling of I/O issues with remote directories. No more stacking dialogs indicating error condition accessing path each time "update grid stats" runs. 2. Alerting when a configured path becomes inaccessible. (assumes stalled process detection is on). Geared toward persons using remote directories. 3. "Suspend Automation" (found in the "process control" main menu). This is useful if you are making configuration changes and don't want any automation tasks to fire. 4. Moved all automation configuration into the "Automation" tab on the configuration editor. Get from SkyDrive [url]http://sdrv.ms/QsaP9Y[/url] |
Thanks!
|
Another successful stall notification. All seems to be working well.
|
1.6.2 on skydrive.
11/04/2012
Changes to 1.6.2 1. Changed GIOM results uploader so it does not force a login for every file being uploaded in a multi-file upload set. After the first login is confirmed all uploads are executed within the same web session. 2. Exported results will now be reordered if required so that the first row in the result set is a "no factor for" row. If the content does not contain any "no factor for" rows the export is cancelled. {this is required to compensate for bug in GIMPS where credits are not given if the first row indicates a factor was found. [B]Information provided by kladner and flashjh[/B]} Get from SkyDrive [URL]http://sdrv.ms/QsaP9Y[/URL] |
good work again Scott !
thank you. Norman |
1.6.4
1.6.4
1.Ability to send email alert when remaining GHZDayToDo drops below threshold 2. On gridStatsUpdate the footer row is always made visible (scrolled to) to keep it in view. This is useful if you have more than 15 instances at which time the scroll-bars become active. [I could not find a way to freeze the bottom row where it was always visible no matter where the scroll position is. I can do it for top rows, but not bottom row] Get from SkyDrive [url]http://sdrv.ms/QsaP9Y[/url] |
It seems I get a stall alert when mfaktx isn't running?
|
[QUOTE=kracker;317470]It seems I get a stall alert when mfaktx isn't running?[/QUOTE]
Are you sure it is not running? I not smart enough to write a program that becomes self aware!:no: Verify it is not in the systray or showing taskManager after you exit. Keep me posted. thx |
Misunderstood.
1 Attachment(s)
[QUOTE=kracker;317470]It seems I get a stall alert when mfaktx isn't running?[/QUOTE]
mfaktXapp cannot determine why the checkpoint files are not changing and therefore you will get a stall detection if you stop mfaktO/C. The key here is that for GPU farms the mfaktO/C instance are on remote machines. Use the Suspend Automation feature or stop mfaktXapp. See screenshot. |
1 Attachment(s)
Thanks, I understand :smile:
[stupid_rant] Yay, day's are getting colder, 5C outside now, time to overcloooock and maybe push it a bit :D Ok, the only reason I'm in the top 5 in the team is because many have "dropped for the season" or many other... the only reason I'm there with my pathetic firepower... [/stupid_rant] |
Kracker are you OK?
[QUOTE=kracker;317594]Thanks, I understand :smile:
[stupid_rant] Yay, day's are getting colder, 5C outside now, time to overcloooock and maybe push it a bit :D Ok, the only reason I'm in the top 5 in the team is because many have "dropped for the season" or many other... the only reason I'm there with my pathetic firepower... [/stupid_rant][/QUOTE] A few too many rum and cokes? |
Dude! Your firepower looks pretty respectable to me. How many cores?! While the GPUs deliver tremendous throughput, when I am at full strength two thirds of my hex core CPU is dedicated to driving mfaktc. Even OC'd to 3.6 GHz on the 1090T, it really takes four cores to saturate the GTX 570. I'm using three right now on the GTX 460.
I lucked out on the 460, catching it as a close out at MicroCenter for just over $100 after rebate. The 570 was really a self-indulgent splurge, and cost about twice as much on Ebay. (I have to note that I also had to upgrade the PSU to run two GPUs, swap in more powerful case and CPU fans, and cut a 120mm hole in the left side panel to mount a third 120mm x 38mm Ultra-Kaze to blow on the GPUs.) Cut yourself some slack, Jack!:razz: [QUOTE=kracker;317594]Thanks, I understand :smile: [stupid_rant] Yay, day's are getting colder, 5C outside now, time to overcloooock and maybe push it a bit :D Ok, the only reason I'm in the top 5 in the team is because many have "dropped for the season" or many other... the only reason I'm there with my pathetic firepower... [/stupid_rant][/QUOTE] |
[QUOTE=kladner;317617]Dude! Your firepower looks pretty respectable to me. How many cores?! While the GPUs deliver tremendous throughput, when I am at full strength two thirds of my hex core CPU is dedicated to driving mfaktc. Even OC'd to 3.6 GHz on the 1090T, it really takes four cores to saturate the GTX 570. I'm using three right now on the GTX 460.
I lucked out on the 460, catching it as a close out at MicroCenter for just over $100 after rebate. The 570 was really a self-indulgent splurge, and cost about twice as much on Ebay. (I have to note that I also had to upgrade the PSU to run two GPUs, swap in more powerful case and CPU fans, and cut a 120mm hole in the left side panel to mount a third 120mm x 38mm Ultra-Kaze to blow on the GPUs.) Cut yourself some slack, Jack!:razz:[/QUOTE] Oh come on, they are all low end, I only have 1 gpu and 1 integrated... anyways my cpu's are 5 Pentium(R) Dual-Core CPU E6550 2 Pentium(R) Dual-Core CPU E5200 2 AMD Dual Core E-300 APU 1 AMD Sempron LE-1300 1 Intel(R) Core(TM) i5-2450M 1 Intel(R) Pentium(R) Dual CPU T3400 And the one with the single-worth med-end gpu.. a AMD A8-3850 feeding it, quad core, no L3 :bangheadonwall: I think the mobile i5 is strongest, lol if I had maybe one or two IB desktop i5's I bet it will blow them away..... :cmd: [SIZE=1][COLOR=DimGray](Hmm.. that reminds me... maybe for christmas.. :devil:) [SIZE=2][COLOR=Black]EDIT: Sorry swl, my brain hasn't been working that well today, or maybe it never has :no:[/COLOR][/SIZE] [/COLOR][/SIZE] |
I would guess you're right about the i5. Dual core HT-enabled, 2.5-3.1 GHz, right?
My Phenom II does moderately well combined with GPUs, but uses more power than i-5/7s. Further, it takes fewer i-something cores to feed mfaktX on any given GPU. You've put together an interesting farm from disparate elements. Here's a thought- can you estimate what your collection uses at full load? For me, with x6 1090T @ 3.6 GHz, GTX 570, GTX 460, the main box alone draws about 670 watts, according to a Kill-a-Watt. Right now, in "economy mode", with the 570 on a Fedex truck heading for City of Industry, CA, it's only pulling ~380 watts. This excludes the ~90 watts for the monitor, which is hooked up separately. It is much quieter under the desk. EDIT: 670 watts is 2x P-1 on 2-3.6 GHz CPU cores, 4x mfaktc on GTX 570 @ 810 MHz, CUDALucas on GTX 460 @ 810 MHz |
[QUOTE=kladner;317626]I would guess you're right about the i5. Dual core HT-enabled, 2.5-3.1 GHz, right?
My Phenom II does moderately well combined with GPUs, but uses more power than i-5/7s. Further, it takes fewer i-something cores to feed mfaktX on any given GPU. You've put together an interesting farm from disparate elements. Here's a thought- can you estimate what your collection uses at full load? For me, with x6 1090T @ 3.6 GHz, GTX 570, GTX 460, the main box alone draws about 670 watts, according to a Kill-a-Watt. Right now, in "economy mode", with the 570 on a Fedex truck heading for City of Industry, CA, it's only pulling ~380 watts. This excludes the ~90 watts for the monitor, which is hooked up separately. It is much quieter under the desk. EDIT: 670 watts is 2x P-1 on 2-3.6 GHz CPU cores, 4x mfaktc on GTX 570 @ 810 MHz, CUDALucas on GTX 460 @ 810 MHz[/QUOTE] I really can't say... I don't have a device to measure power consumption, however, I believe most of them don't have a TDP over 50W (except for the one which powers my gpu) |
You might be getting a pretty good return there, then. Hard to estimate, of course. I did not do as well as I might have when I built the foundation of my one current system.
EDIT: The GPUs came later. I was thinking more Photoshop than GIMPS when I started out. |
[QUOTE=kladner;317631]You might be getting a pretty good return there, then. Hard to estimate, of course. I did not do as well as I might have when I built the foundation of my one current system.[/QUOTE]
Yeah, I got the core2 duo's just recently, I'm hoping I can get 1 DC a day with everything, but I'm doubtful... well I don't know yet since they're not all on (yet) |
can some supermod move the last posts in the related hardware subforum, and maybe rename the thread title here to take out the "1.4", as I see the current version is now, what? 1.6?
|
[QUOTE=LaurV;317633]can some supermod move the last posts in the related hardware subforum, and maybe rename the thread title here to take out the "1.4", as I see the current version is now, what? 1.6?[/QUOTE]
Sorry LaurV. Point taken.:ermm: |
[QUOTE=LaurV;317633]can some supermod move the last posts in the related hardware subforum, and maybe rename the thread title here to take out the "1.4", as I see the current version is now, what? 1.6?[/QUOTE]
*sniff* :cry: Yes mi'lord, I understand. I'll stop. :cmd: |
| All times are UTC. The time now is 08:27. |
Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.