Wassup is I gathered info from all sites I could find on what identifies when a new core was used for a CPU, but these various sites aren't terribly detailed in what they say (and often conflict with one another). Not to mention that AMD and Intel don't even publish a document with any way to identify cores that I could find.
So...if you provide the entire string of its mis-identification, (eg: AMD K8, Athlon 64 bla blah, xxxMhz (Clawhammer nnn-nn)) I can use the debugging info to change the ID table. ;)
> Wassup :)
>
> Wassup is I gathered info from all sites I could find on what identifies when
> a new core was used for a CPU, but these various sites aren't terribly
> detailed in what they say (and often conflict with one another). Not to
> mention that AMD and Intel don't even publish a document with any way to
> identify cores that I could find.
>
> So...if you provide the entire string of its mis-identification, (eg:
> AMD K8, Athlon 64 bla blah, xxxMhz (Clawhammer nnn-nn)) I can use the
> debugging info to change the ID table. ;)
>
>
While you seem to be coding away, is there an option to display what switches you specify on program launch in the messages window?
Starting BOINC client version 4.19 for windows_intelx86
Project prefs: using separate prefs for home
Project prefs: no separate prefs for home; using your defaults
Host ID is 61377
Host ID is not assigned yet
General prefs: from Einstein@Home (last modified 2005-03-11 00:29:33)
General prefs: no separate prefs for home; using your defaults
CPU: AMD K8, Athlon-64, 2066Mhz, (Newcastle-FC0-FC0)
CPU: Cache L1=128k, L2=512k, L3=0k
CPU: count=1, HT flag=no, MP flag=no (0)
Flags: -return_results_immediately -min
> Sure,
>
> I'll call it [BOINC Cmd line: "-return_results_immediately -min" ]
>
> Its on the update site... href="https://einsteinathome.org/%3Ca%20href%3D"http://www.freewebs.com/gladry/boinc_gui.exeFile">http://www.freewebs.com/gladry/boinc_gui.exeFile">boinc_gui.exeFile[/url]
>
> Also the MHZ should be much more accurate (we hope).
>
>
> Sure,
>
> I'll call it [BOINC Cmd line: "-return_results_immediately -min" ]
>
> Its on the update site... href="https://einsteinathome.org/%3Ca%20href%3D"http://www.freewebs.com/gladry/boinc_gui.exeFile">http://www.freewebs.com/gladry/boinc_gui.exeFile">boinc_gui.exeFile[/url]
>
> Also the MHZ should be much more accurate (we hope).
>
>
The previous output said 2066Mhz, that wasn't an error in your code, it was because my Gigabyte K8VT800 Pro comes with a nify processor tool, which lets me overclock it :) So if you thought the code was wrong because of my posted output... it wasn't :S Because the app now reads my processor as a 2000Mhz again.
Do you know of any way to make BOINC write to the registry, in such as way as this:
Slots would contain (in numerical form) the current number of slots Boinc is using, like Einstein+Seti@Home+Pirate@Home = 3
Slot_?.Name would contain the Project Name for each slot, Slot_?.Percentage would contain the progress of each work unit.
Slot_?.EstCompletion would contain the time left until completion of the work unit.
Each string in the key would have to be assigned a value of 0 when the program is started, and would then write new statistics afterwards.
I tried this myself, but I got loads of build errors in Visual Studio 2005 Beta 2 (Whidbey), and they were all in parts of the code I didn't modify :/
The hole point of the above registry work, is so I can make a DesktopX object to monitor my Boinc :D Which I would gladly post... if I ever got it working.
Thanks, and sorry to keep asking so much of you :S
PS: Boinc Cmd line: "Files\Boinc\Boinc_gui.exe" -return_results_immediately" keeps appearing, now, which is good :) Is my shortcut making "Files\Boinc\Boinc_gui.exe" appear?? Or is it possible to mod the code, so that it only displays the switches, rather than the path too?
> Do you know of any way to make BOINC write to the registry, in such as way as
> this:
>
I believe all that information is contained inside of "client_state.xml".
Even if it isnt, the method to get those numbers is by using something called "RPC" - Remote Procedure Calls. Basically they are an ethernet messaging system.
"Hey boinc, what apps you runnining..."
"Hey boinc, pause this WU"
"Hey boinc, what are the messages in your message window"
And so forth.
For an example of a program using them, try boincview.
> You might want to try a modified windows 4.19 BOINC client I'm using.
>
> It has a few benchmarking features and a few others as well.
I have finally gotten irritated enough by the erratic "requested credits" by BOINC to try your version (running under win2000). I un-installed BOINC 4.25 & re-installed 4.19 first.
Your version did give more detailed information about the CPUs (dual AthlonMPs). I don't think it solved the annoying "requested credits" problem with dual CPUs though.
Since the requested credits are based on benchmark calculations, I looked at the reported benchmarks. I am going to try a table & hope this website does not totally destroy my table formatting.
BENCHMARKING THE BENCHMARKS
(all on the same AthlonMP computer)
The variations in benchmarks are rather odd, but I won't comment further because they don't seem big enough to account for the requested credits problem. I am more inclined to suspect that there is a factor-of-two (2 CPUs) problem somewhere in the BOINC claimed credits calculation.
I am convinced that nearly all E@H units distributed during this time had virtually the same computing power requirement & the median claimed & awarded credit for these was in the ballpark of 80 points. Any variation is just strange behavior by BOINC, unrelated to the real units.
I would like to ask (of ben or anyone else who might stray into this thread) if there is any way I might obtain the sourcecode to BOINC 4.19 or 4.25.
Wassup :) Wassup is I
)
Wassup :)
Wassup is I gathered info from all sites I could find on what identifies when a new core was used for a CPU, but these various sites aren't terribly detailed in what they say (and often conflict with one another). Not to mention that AMD and Intel don't even publish a document with any way to identify cores that I could find.
So...if you provide the entire string of its mis-identification, (eg: AMD K8, Athlon 64 bla blah, xxxMhz (Clawhammer nnn-nn)) I can use the debugging info to change the ID table. ;)
> Wassup :) > > Wassup is I
)
> Wassup :)
>
> Wassup is I gathered info from all sites I could find on what identifies when
> a new core was used for a CPU, but these various sites aren't terribly
> detailed in what they say (and often conflict with one another). Not to
> mention that AMD and Intel don't even publish a document with any way to
> identify cores that I could find.
>
> So...if you provide the entire string of its mis-identification, (eg:
> AMD K8, Athlon 64 bla blah, xxxMhz (Clawhammer nnn-nn)) I can use the
> debugging info to change the ID table. ;)
>
>
AMD, K8, Athlon-64, 2000 Mhz, (Clawhammer-FC0-FC0)
Everything is fine, until you get to the clawhammer part :)
> > AMD, K8, Athlon-64, 2000
)
>
> AMD, K8, Athlon-64, 2000 Mhz, (Clawhammer-FC0-FC0)
>
> Everything is fine, until you get to the clawhammer part :)
New version posted.
boinc_gui.exeFile
if(cpu.family == 15) {
{15,99,0,4,0x00<<6,-1,"Athlon-64","K8","Engineering"},
{15,99,0,4,0x04<<6,-1,"Athlon-64","K8","Clawhammer"},
{15,99,0,4,0x08<<6,-1,"Athlon-64 Mobile","K8","Clawhammer"},
{15,99,0,5,0x0C<<6,-1,"Opteron 1","K8o","Sledgehammer"},
{15,99,0,5,0x10<<6,-1,"Opteron 2","K8o","Sledgehammer"},
{15,99,0,5,0x14<<6,-1,"Opteron 8","K8o","Sledgehammer"},
{15,99,0,5,0x1D<<6,-1,"Athlon-64 XP-M","K8","Sledgehammer"},
{15,99,0,5,0x20<<6,-1,"Athlon XP","K8","Sledgehammer"},
{15,99,0,5,0x21<<6,-1,"Sempron M","K8","Newcastle"},
{15,99,0,5,0x22<<6,-1,"Sempron","K8","Newcastle"},
{15,99,0,5,0x23<<6,-1,"Sempron M","K8","Newcastle"},
{15,99,0,5,0x24<<6,-1,"Athlon-64 FX","K8x","Sledgehammer"},
if(cache.L2 < 1024) name.nick = "Newcastle";
if(cache.L2
> > > > AMD, K8, Athlon-64,
)
> >
> > AMD, K8, Athlon-64, 2000 Mhz, (Clawhammer-FC0-FC0)
> >
> > Everything is fine, until you get to the clawhammer part :)
>
> New version posted.
>
> href="https://einsteinathome.org/%3Ca%20href%3D"http://www.freewebs.com/gladry/boinc_gui.exeFile">http://www.freewebs.com/gladry/boinc_gui.exeFile">boinc_gui.exeFile[/url]
>
> if(cpu.family == 15) {
> {15,99,0,4,0x00 {15,99,0,4,0x04 {15,99,0,4,0x08 Mobile","K8","Clawhammer"},
> {15,99,0,5,0x0C {15,99,0,5,0x10 {15,99,0,5,0x14 {15,99,0,5,0x1D XP-M","K8","Sledgehammer"},
> {15,99,0,5,0x20 {15,99,0,5,0x21 {15,99,0,5,0x22 {15,99,0,5,0x23 {15,99,0,5,0x24
> if(cache.L2
>
> if(cache.L2
>
Nice one :) Mirrored again at http://www.jamieplucinski.co.uk/downloads/boinc_gui.exe :)
While you seem to be coding
)
While you seem to be coding away, is there an option to display what switches you specify on program launch in the messages window?
Starting BOINC client version 4.19 for windows_intelx86
Project prefs: using separate prefs for home
Project prefs: no separate prefs for home; using your defaults
Host ID is 61377
Host ID is not assigned yet
General prefs: from Einstein@Home (last modified 2005-03-11 00:29:33)
General prefs: no separate prefs for home; using your defaults
CPU: AMD K8, Athlon-64, 2066Mhz, (Newcastle-FC0-FC0)
CPU: Cache L1=128k, L2=512k, L3=0k
CPU: count=1, HT flag=no, MP flag=no (0)
Flags: -return_results_immediately -min
Is it a possibility?? :D
Sure, I'll call it [BOINC
)
Sure,
I'll call it [BOINC Cmd line: "-return_results_immediately -min" ]
Its on the update site...boinc_gui.exeFile
Also the MHZ should be much more accurate (we hope).
> Sure, > > I'll call it
)
> Sure,
>
> I'll call it [BOINC Cmd line: "-return_results_immediately -min" ]
>
> Its on the update site... href="https://einsteinathome.org/%3Ca%20href%3D"http://www.freewebs.com/gladry/boinc_gui.exeFile">http://www.freewebs.com/gladry/boinc_gui.exeFile">boinc_gui.exeFile[/url]
>
> Also the MHZ should be much more accurate (we hope).
>
>
Thanks :D
> Sure, > > I'll call it
)
> Sure,
>
> I'll call it [BOINC Cmd line: "-return_results_immediately -min" ]
>
> Its on the update site... href="https://einsteinathome.org/%3Ca%20href%3D"http://www.freewebs.com/gladry/boinc_gui.exeFile">http://www.freewebs.com/gladry/boinc_gui.exeFile">boinc_gui.exeFile[/url]
>
> Also the MHZ should be much more accurate (we hope).
>
>
The previous output said 2066Mhz, that wasn't an error in your code, it was because my Gigabyte K8VT800 Pro comes with a nify processor tool, which lets me overclock it :) So if you thought the code was wrong because of my posted output... it wasn't :S Because the app now reads my processor as a 2000Mhz again.
Do you know of any way to make BOINC write to the registry, in such as way as this:
Key: HKEY_LOCAL_MACHINE\Software\Boinc\Work
String Values Contained Within Key:
Slots
Slot_1.Name
Slot_1.Percentage
Slot_1.EstCompletion
Slot_2.Name
Slot_2.Percentage
Slot_2.EstCompletion
Slot_3.Name
Slot_3.Percentage
Slot_3.EstCompletion
Slot_4.Name
Slot_4.Percentage
Slot_4.EstCompletion
...etc
Slots would contain (in numerical form) the current number of slots Boinc is using, like Einstein+Seti@Home+Pirate@Home = 3
Slot_?.Name would contain the Project Name for each slot, Slot_?.Percentage would contain the progress of each work unit.
Slot_?.EstCompletion would contain the time left until completion of the work unit.
Each string in the key would have to be assigned a value of 0 when the program is started, and would then write new statistics afterwards.
I tried this myself, but I got loads of build errors in Visual Studio 2005 Beta 2 (Whidbey), and they were all in parts of the code I didn't modify :/
The hole point of the above registry work, is so I can make a DesktopX object to monitor my Boinc :D Which I would gladly post... if I ever got it working.
Thanks, and sorry to keep asking so much of you :S
PS: Boinc Cmd line: "Files\Boinc\Boinc_gui.exe" -return_results_immediately" keeps appearing, now, which is good :) Is my shortcut making "Files\Boinc\Boinc_gui.exe" appear?? Or is it possible to mod the code, so that it only displays the switches, rather than the path too?
> Do you know of any way to
)
> Do you know of any way to make BOINC write to the registry, in such as way as
> this:
>
I believe all that information is contained inside of "client_state.xml".
Even if it isnt, the method to get those numbers is by using something called "RPC" - Remote Procedure Calls. Basically they are an ethernet messaging system.
"Hey boinc, what apps you runnining..."
"Hey boinc, pause this WU"
"Hey boinc, what are the messages in your message window"
And so forth.
For an example of a program using them, try boincview.
ben wrote: > You might
)
ben wrote:
> You might want to try a modified windows 4.19 BOINC client I'm using.
>
> It has a few benchmarking features and a few others as well.
I have finally gotten irritated enough by the erratic "requested credits" by BOINC to try your version (running under win2000). I un-installed BOINC 4.25 & re-installed 4.19 first.
Your version did give more detailed information about the CPUs (dual AthlonMPs). I don't think it solved the annoying "requested credits" problem with dual CPUs though.
Since the requested credits are based on benchmark calculations, I looked at the reported benchmarks. I am going to try a table & hope this website does not totally destroy my table formatting.
BENCHMARKING THE BENCHMARKS
(all on the same AthlonMP computer)
Date--- BOINC-- operating--- benchmarks--- requested
------- version system------ float - int-- credit
------- -------- ------------- ---- -------- ---------
2/25/05 4.19---- native linux 950 1992--- ?
3/03/05 4.19---- wine-linux 1676 1397--- ?
3/09/05 4.19---- wine-linux 1676 1439--- ?
3/11/05 4.25---- win2000---- 1684 1193--- 43.3
4/02/05 Ben's*-- win2000---- 1184 1196**. 46.2
4/02/05 Ben's*-- win2000---- 1537 1537*** 46.2
* Ben's modification of 4.19.
** printed by BOINC messages
*** posted in E@H web site
The variations in benchmarks are rather odd, but I won't comment further because they don't seem big enough to account for the requested credits problem. I am more inclined to suspect that there is a factor-of-two (2 CPUs) problem somewhere in the BOINC claimed credits calculation.
I am convinced that nearly all E@H units distributed during this time had virtually the same computing power requirement & the median claimed & awarded credit for these was in the ballpark of 80 points. Any variation is just strange behavior by BOINC, unrelated to the real units.
I would like to ask (of ben or anyone else who might stray into this thread) if there is any way I might obtain the sourcecode to BOINC 4.19 or 4.25.
Tnanks,
ADDMP