Re: [OMPI devel] ob1 and req->req_state
On Jun 23, 2008, at 5:51 PM, Jeff Squyres wrote: Ah -- I see -- we have 2 different fields with the same name (just different places within the struct hierarchy) with different meanings. That was a good idea. ;-) exactly Thanks; that actually helps understand things quite a bit. On Jun 23, 2008, at 5:45 PM, Shipman, Galen M. wrote: Oh, I see, you are confusing the req_state on the OMPI request with the req_state on the PML request. The ompi request state is for persistent requests, the PML request state is not and does not use that enum. - Galen On Jun 23, 2008, at 5:18 PM, Jeff Squyres wrote: On Jun 23, 2008, at 4:43 PM, Shipman, Galen M. wrote: We use req_state currently to track that we receive both RNDV completion and RNDV ack prior to freeing the request.. Does that mean you're not using the enum values, but rather just to indicate that the value is >= 0? -- Jeff Squyres Cisco Systems ___ devel mailing list de...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/devel ___ devel mailing list de...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/devel -- Jeff Squyres Cisco Systems ___ devel mailing list de...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/devel
Re: [OMPI devel] ob1 and req->req_state
Ah -- I see -- we have 2 different fields with the same name (just different places within the struct hierarchy) with different meanings. That was a good idea. ;-) Thanks; that actually helps understand things quite a bit. On Jun 23, 2008, at 5:45 PM, Shipman, Galen M. wrote: Oh, I see, you are confusing the req_state on the OMPI request with the req_state on the PML request. The ompi request state is for persistent requests, the PML request state is not and does not use that enum. - Galen On Jun 23, 2008, at 5:18 PM, Jeff Squyres wrote: On Jun 23, 2008, at 4:43 PM, Shipman, Galen M. wrote: We use req_state currently to track that we receive both RNDV completion and RNDV ack prior to freeing the request.. Does that mean you're not using the enum values, but rather just to indicate that the value is >= 0? -- Jeff Squyres Cisco Systems ___ devel mailing list de...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/devel ___ devel mailing list de...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/devel -- Jeff Squyres Cisco Systems
Re: [OMPI devel] ob1 and req->req_state
Oh, I see, you are confusing the req_state on the OMPI request with the req_state on the PML request. The ompi request state is for persistent requests, the PML request state is not and does not use that enum. - Galen On Jun 23, 2008, at 5:18 PM, Jeff Squyres wrote: On Jun 23, 2008, at 4:43 PM, Shipman, Galen M. wrote: We use req_state currently to track that we receive both RNDV completion and RNDV ack prior to freeing the request.. Does that mean you're not using the enum values, but rather just to indicate that the value is >= 0? -- Jeff Squyres Cisco Systems ___ devel mailing list de...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/devel
Re: [OMPI devel] PML selection logic
Okay, so let's explore an alternative that preserves the support you are seeking for the "ignorant user", but doesn't penalize everyone else. What we could do is simply set things up so that: 1. if -mca plm xyz is provided, then no modex data is added 2. if it is not provided, then only rank=0 inserts the data. All other procs simply check their own selection against the one given by rank=0 Now, if a knowledgeable user or sys admin specifies what to use for their system, we won't penalize their startup time. A user who doesn't know what to do gets to run, albeit less scalably on startup. Looking forward from there, we can look to a day where failing to initialize something that exists on the system could be detected in some other fashion, letting the local proc abort since it would know that other procs that detected similar capabilities may well have selected that PML. For now, though, this would solve the problem. Make sense? Ralph On 6/23/08 1:31 PM, "Brian W. Barrett"wrote: > The problem is that we default to OB1, but that's not the right choice for > some platforms (like Pathscale / PSM), where there's a huge performance > hit for using OB1. So we run into a situation where user installs Open > MPI, starts running, gets horrible performance, bad mouths Open MPI, and > now we're in that game again. Yeah, the sys admin should know what to do, > but it doesn't always work that way. > > Brian > > > On Mon, 23 Jun 2008, Ralph H Castain wrote: > >> My fault - I should be more precise in my language. ;-/ >> >> #1 is not adequate, IMHO, as it forces us to -always- do a modex. It seems >> to me that a simpler solution to what you describe is for the user to >> specify -mca pml ob1, or -mca pml cm. If the latter, then you could deal >> with the failed-to-initialize problem cleanly by having the proc directly >> abort. >> >> Again, sometimes I think we attempt to automate too many things. This seems >> like a pretty clear case where you know what you want - the sys admin, if >> nobody else, can certainly set that mca param in the default param file! >> >> Otherwise, it seems to me that you are relying on the modex to detect that >> your proc failed to init the correct subsystem. I hate to force a modex just >> for that - if so, then perhaps this could again be a settable option to >> avoid requiring non-scalable behavior for those of us who want scalability? >> >> >> On 6/23/08 1:21 PM, "Brian W. Barrett" wrote: >> >>> The selection code was added because frequently high speed interconnects >>> fail to initialize properly due to random stuff happening (yes, that's a >>> horrible statement, but true). We ran into a situation with some really >>> flaky machines where most of the processes would chose CM, but a couple >>> would fail to initialize the MTL and therefore chose OB1. This lead to a >>> hang situation, which is the worst of the worst. >>> >>> I think #1 is adequate, although it doesn't handle spawn particularly >>> well. And spawn is generally used in environments where such network >>> mismatches are most likely to occur. >>> >>> Brian >>> >>> >>> On Mon, 23 Jun 2008, Ralph H Castain wrote: >>> Since my goal is to eliminate the modex completely for managed installations, could you give me a brief understanding of this eventual PML selection logic? It would help to hear an example of how and why different procs could get different answers - and why we would want to allow them to do so. Thanks Ralph On 6/23/08 11:59 AM, "Aurélien Bouteiller" wrote: > The first approach sounds fair enough to me. We should avoid 2 and 3 > as the pml selection mechanism used to be > more complex before we reduced it to accommodate a major design bug in > the BTL selection process. When using the complete PML selection, BTL > would be initialized several times, leading to a variety of bugs. > Eventually the PML selection should return to its old self, when the > BTL bug gets fixed. > > Aurelien > > Le 23 juin 08 à 12:36, Ralph H Castain a écrit : > >> Yo all >> >> I've been doing further research into the modex and came across >> something I >> don't fully understand. It seems we have each process insert into >> the modex >> the name of the PML module that it selected. Once the modex has >> exchanged >> that info, it then loops across all procs in the job to check their >> selection, and aborts if any proc picked a different PML module. >> >> All well and good...assuming that procs actually -can- choose >> different PML >> modules and hence create an "abort" scenario. However, if I look >> inside the >> PML's at their selection logic, I find that a proc can ONLY pick a >> module >> other than ob1 if: >> >> 1. the user
Re: [OMPI devel] PML selection logic
The problem is that we default to OB1, but that's not the right choice for some platforms (like Pathscale / PSM), where there's a huge performance hit for using OB1. So we run into a situation where user installs Open MPI, starts running, gets horrible performance, bad mouths Open MPI, and now we're in that game again. Yeah, the sys admin should know what to do, but it doesn't always work that way. Brian On Mon, 23 Jun 2008, Ralph H Castain wrote: My fault - I should be more precise in my language. ;-/ #1 is not adequate, IMHO, as it forces us to -always- do a modex. It seems to me that a simpler solution to what you describe is for the user to specify -mca pml ob1, or -mca pml cm. If the latter, then you could deal with the failed-to-initialize problem cleanly by having the proc directly abort. Again, sometimes I think we attempt to automate too many things. This seems like a pretty clear case where you know what you want - the sys admin, if nobody else, can certainly set that mca param in the default param file! Otherwise, it seems to me that you are relying on the modex to detect that your proc failed to init the correct subsystem. I hate to force a modex just for that - if so, then perhaps this could again be a settable option to avoid requiring non-scalable behavior for those of us who want scalability? On 6/23/08 1:21 PM, "Brian W. Barrett"wrote: The selection code was added because frequently high speed interconnects fail to initialize properly due to random stuff happening (yes, that's a horrible statement, but true). We ran into a situation with some really flaky machines where most of the processes would chose CM, but a couple would fail to initialize the MTL and therefore chose OB1. This lead to a hang situation, which is the worst of the worst. I think #1 is adequate, although it doesn't handle spawn particularly well. And spawn is generally used in environments where such network mismatches are most likely to occur. Brian On Mon, 23 Jun 2008, Ralph H Castain wrote: Since my goal is to eliminate the modex completely for managed installations, could you give me a brief understanding of this eventual PML selection logic? It would help to hear an example of how and why different procs could get different answers - and why we would want to allow them to do so. Thanks Ralph On 6/23/08 11:59 AM, "Aurélien Bouteiller" wrote: The first approach sounds fair enough to me. We should avoid 2 and 3 as the pml selection mechanism used to be more complex before we reduced it to accommodate a major design bug in the BTL selection process. When using the complete PML selection, BTL would be initialized several times, leading to a variety of bugs. Eventually the PML selection should return to its old self, when the BTL bug gets fixed. Aurelien Le 23 juin 08 à 12:36, Ralph H Castain a écrit : Yo all I've been doing further research into the modex and came across something I don't fully understand. It seems we have each process insert into the modex the name of the PML module that it selected. Once the modex has exchanged that info, it then loops across all procs in the job to check their selection, and aborts if any proc picked a different PML module. All well and good...assuming that procs actually -can- choose different PML modules and hence create an "abort" scenario. However, if I look inside the PML's at their selection logic, I find that a proc can ONLY pick a module other than ob1 if: 1. the user specifies the module to use via -mca pml xyz or by using a module specific mca param to adjust its priority. In this case, since the mca param is propagated, ALL procs have no choice but to pick that same module, so that can't cause us to abort (we will have already returned an error and aborted if the specified module can't run). 2. the pml/cm module detects that an MTL module was selected, and that it is other than "psm". In this case, the CM module will be selected because its default priority is higher than that of OB1. In looking deeper into the MTL selection logic, it appears to me that you either have the required capability or you don't. I can see that in some environments (e.g., rsh across unmanaged collections of machines), it might be possible for someone to launch across a set of machines where some do and some don't have the required support. However, in all other cases, this will be homogeneous across the system. Given this analysis (and someone more familiar with the PML should feel free to confirm or correct it), it seems to me that this could be streamlined via one or more means: 1. at the most, we could have rank=0 add the PML module name to the modex, and other procs simply check it against their own and return an error if they differ. This accomplishes the identical functionality to what we have today, but with much less info in the modex. 2. we could eliminate this info from the modex altogether by
Re: [OMPI devel] PML selection logic
My fault - I should be more precise in my language. ;-/ #1 is not adequate, IMHO, as it forces us to -always- do a modex. It seems to me that a simpler solution to what you describe is for the user to specify -mca pml ob1, or -mca pml cm. If the latter, then you could deal with the failed-to-initialize problem cleanly by having the proc directly abort. Again, sometimes I think we attempt to automate too many things. This seems like a pretty clear case where you know what you want - the sys admin, if nobody else, can certainly set that mca param in the default param file! Otherwise, it seems to me that you are relying on the modex to detect that your proc failed to init the correct subsystem. I hate to force a modex just for that - if so, then perhaps this could again be a settable option to avoid requiring non-scalable behavior for those of us who want scalability? On 6/23/08 1:21 PM, "Brian W. Barrett"wrote: > The selection code was added because frequently high speed interconnects > fail to initialize properly due to random stuff happening (yes, that's a > horrible statement, but true). We ran into a situation with some really > flaky machines where most of the processes would chose CM, but a couple > would fail to initialize the MTL and therefore chose OB1. This lead to a > hang situation, which is the worst of the worst. > > I think #1 is adequate, although it doesn't handle spawn particularly > well. And spawn is generally used in environments where such network > mismatches are most likely to occur. > > Brian > > > On Mon, 23 Jun 2008, Ralph H Castain wrote: > >> Since my goal is to eliminate the modex completely for managed >> installations, could you give me a brief understanding of this eventual PML >> selection logic? It would help to hear an example of how and why different >> procs could get different answers - and why we would want to allow them to >> do so. >> >> Thanks >> Ralph >> >> >> >> On 6/23/08 11:59 AM, "Aurélien Bouteiller" wrote: >> >>> The first approach sounds fair enough to me. We should avoid 2 and 3 >>> as the pml selection mechanism used to be >>> more complex before we reduced it to accommodate a major design bug in >>> the BTL selection process. When using the complete PML selection, BTL >>> would be initialized several times, leading to a variety of bugs. >>> Eventually the PML selection should return to its old self, when the >>> BTL bug gets fixed. >>> >>> Aurelien >>> >>> Le 23 juin 08 à 12:36, Ralph H Castain a écrit : >>> Yo all I've been doing further research into the modex and came across something I don't fully understand. It seems we have each process insert into the modex the name of the PML module that it selected. Once the modex has exchanged that info, it then loops across all procs in the job to check their selection, and aborts if any proc picked a different PML module. All well and good...assuming that procs actually -can- choose different PML modules and hence create an "abort" scenario. However, if I look inside the PML's at their selection logic, I find that a proc can ONLY pick a module other than ob1 if: 1. the user specifies the module to use via -mca pml xyz or by using a module specific mca param to adjust its priority. In this case, since the mca param is propagated, ALL procs have no choice but to pick that same module, so that can't cause us to abort (we will have already returned an error and aborted if the specified module can't run). 2. the pml/cm module detects that an MTL module was selected, and that it is other than "psm". In this case, the CM module will be selected because its default priority is higher than that of OB1. In looking deeper into the MTL selection logic, it appears to me that you either have the required capability or you don't. I can see that in some environments (e.g., rsh across unmanaged collections of machines), it might be possible for someone to launch across a set of machines where some do and some don't have the required support. However, in all other cases, this will be homogeneous across the system. Given this analysis (and someone more familiar with the PML should feel free to confirm or correct it), it seems to me that this could be streamlined via one or more means: 1. at the most, we could have rank=0 add the PML module name to the modex, and other procs simply check it against their own and return an error if they differ. This accomplishes the identical functionality to what we have today, but with much less info in the modex. 2. we could eliminate this info from the modex altogether by requiring the user to specify the
Re: [OMPI devel] ob1 and req->req_state
On Mon, 23 Jun 2008, Jeff Squyres wrote: On Jun 23, 2008, at 3:17 PM, Brian W. Barrett wrote: Just because it's volatile doesn't mean that adds are atomic. There's at least one place in the PML (or used to be) where two threads could decrement that counter at the same time. With atomics, then both subtracts should occur, right? So a request could go from ACTIVE -> INACTIVE -> INVALID. Is that what is desired? (I honestly don't know enough about ob1 to say) Or should we just be assigning a specific state, rather than relying on subtracting? That was my real question. I honestly don't know. I just remember that there were some cases where we were doing crazy counting. Brian
Re: [OMPI devel] PML selection logic
The selection code was added because frequently high speed interconnects fail to initialize properly due to random stuff happening (yes, that's a horrible statement, but true). We ran into a situation with some really flaky machines where most of the processes would chose CM, but a couple would fail to initialize the MTL and therefore chose OB1. This lead to a hang situation, which is the worst of the worst. I think #1 is adequate, although it doesn't handle spawn particularly well. And spawn is generally used in environments where such network mismatches are most likely to occur. Brian On Mon, 23 Jun 2008, Ralph H Castain wrote: Since my goal is to eliminate the modex completely for managed installations, could you give me a brief understanding of this eventual PML selection logic? It would help to hear an example of how and why different procs could get different answers - and why we would want to allow them to do so. Thanks Ralph On 6/23/08 11:59 AM, "Aurélien Bouteiller"wrote: The first approach sounds fair enough to me. We should avoid 2 and 3 as the pml selection mechanism used to be more complex before we reduced it to accommodate a major design bug in the BTL selection process. When using the complete PML selection, BTL would be initialized several times, leading to a variety of bugs. Eventually the PML selection should return to its old self, when the BTL bug gets fixed. Aurelien Le 23 juin 08 à 12:36, Ralph H Castain a écrit : Yo all I've been doing further research into the modex and came across something I don't fully understand. It seems we have each process insert into the modex the name of the PML module that it selected. Once the modex has exchanged that info, it then loops across all procs in the job to check their selection, and aborts if any proc picked a different PML module. All well and good...assuming that procs actually -can- choose different PML modules and hence create an "abort" scenario. However, if I look inside the PML's at their selection logic, I find that a proc can ONLY pick a module other than ob1 if: 1. the user specifies the module to use via -mca pml xyz or by using a module specific mca param to adjust its priority. In this case, since the mca param is propagated, ALL procs have no choice but to pick that same module, so that can't cause us to abort (we will have already returned an error and aborted if the specified module can't run). 2. the pml/cm module detects that an MTL module was selected, and that it is other than "psm". In this case, the CM module will be selected because its default priority is higher than that of OB1. In looking deeper into the MTL selection logic, it appears to me that you either have the required capability or you don't. I can see that in some environments (e.g., rsh across unmanaged collections of machines), it might be possible for someone to launch across a set of machines where some do and some don't have the required support. However, in all other cases, this will be homogeneous across the system. Given this analysis (and someone more familiar with the PML should feel free to confirm or correct it), it seems to me that this could be streamlined via one or more means: 1. at the most, we could have rank=0 add the PML module name to the modex, and other procs simply check it against their own and return an error if they differ. This accomplishes the identical functionality to what we have today, but with much less info in the modex. 2. we could eliminate this info from the modex altogether by requiring the user to specify the PML module if they want something other than the default OB1. In this case, there can be no confusion over what each proc is to use. The CM module will attempt to init the MTL - if it cannot do so, then the job will return the correct error and tell the user that CM/MTL support is unavailable. 3. we could again eliminate the info by not inserting it into the modex if (a) the default PML module is selected, or (b) the user specified the PML module to be used. In the first case, each proc can simply check to see if they picked the default - if not, then we can insert the info to indicate the difference. Thus, in the "standard" case, no info will be inserted. In the second case, we will already get an error if the specified PML module could not be used. Hence, the modex check provides no additional info or value. I understand the motivation to support automation. However, in this case, the automation actually doesn't seem to buy us very much, and it isn't coming "free". So perhaps some change in how this is done would be in order? Ralph ___ devel mailing list de...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/devel ___ devel mailing list de...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/devel
Re: [OMPI devel] ob1 and req->req_state
On Jun 23, 2008, at 3:17 PM, Brian W. Barrett wrote: Just because it's volatile doesn't mean that adds are atomic. There's at least one place in the PML (or used to be) where two threads could decrement that counter at the same time. With atomics, then both subtracts should occur, right? So a request could go from ACTIVE -> INACTIVE -> INVALID. Is that what is desired? (I honestly don't know enough about ob1 to say) Or should we just be assigning a specific state, rather than relying on subtracting? That was my real question. On Mon, 23 Jun 2008, Jeff Squyres wrote: I see in a few places in ob1 we do things like this: OPAL_THREAD_ADD32(>req_state, -1); Why do we do this? req_state is technically an enum value, so we shouldn't be adding/subtracting to it (granted, it looks like the enum values were carefully chosen to allow this). Additionally, req_state is volatile; the atomics shouldn't be necessary. Is there some other non-obvious reason? Also, I see this in a few places: req->req_state = 2; which really should be req->req_state = OMPI_REQUEST_ACTIVE; ___ devel mailing list de...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/devel -- Jeff Squyres Cisco Systems
Re: [OMPI devel] ob1 and req->req_state
Just because it's volatile doesn't mean that adds are atomic. There's at least one place in the PML (or used to be) where two threads could decrement that counter at the same time. Brian On Mon, 23 Jun 2008, Jeff Squyres wrote: I see in a few places in ob1 we do things like this: OPAL_THREAD_ADD32(>req_state, -1); Why do we do this? req_state is technically an enum value, so we shouldn't be adding/subtracting to it (granted, it looks like the enum values were carefully chosen to allow this). Additionally, req_state is volatile; the atomics shouldn't be necessary. Is there some other non-obvious reason? Also, I see this in a few places: req->req_state = 2; which really should be req->req_state = OMPI_REQUEST_ACTIVE;
[OMPI devel] ob1 and req->req_state
I see in a few places in ob1 we do things like this: OPAL_THREAD_ADD32(>req_state, -1); Why do we do this? req_state is technically an enum value, so we shouldn't be adding/subtracting to it (granted, it looks like the enum values were carefully chosen to allow this). Additionally, req_state is volatile; the atomics shouldn't be necessary. Is there some other non-obvious reason? Also, I see this in a few places: req->req_state = 2; which really should be req->req_state = OMPI_REQUEST_ACTIVE; -- Jeff Squyres Cisco Systems
Re: [OMPI devel] PML selection logic
The first approach sounds fair enough to me. We should avoid 2 and 3 as the pml selection mechanism used to be more complex before we reduced it to accommodate a major design bug in the BTL selection process. When using the complete PML selection, BTL would be initialized several times, leading to a variety of bugs. Eventually the PML selection should return to its old self, when the BTL bug gets fixed. Aurelien Le 23 juin 08 à 12:36, Ralph H Castain a écrit : Yo all I've been doing further research into the modex and came across something I don't fully understand. It seems we have each process insert into the modex the name of the PML module that it selected. Once the modex has exchanged that info, it then loops across all procs in the job to check their selection, and aborts if any proc picked a different PML module. All well and good...assuming that procs actually -can- choose different PML modules and hence create an "abort" scenario. However, if I look inside the PML's at their selection logic, I find that a proc can ONLY pick a module other than ob1 if: 1. the user specifies the module to use via -mca pml xyz or by using a module specific mca param to adjust its priority. In this case, since the mca param is propagated, ALL procs have no choice but to pick that same module, so that can't cause us to abort (we will have already returned an error and aborted if the specified module can't run). 2. the pml/cm module detects that an MTL module was selected, and that it is other than "psm". In this case, the CM module will be selected because its default priority is higher than that of OB1. In looking deeper into the MTL selection logic, it appears to me that you either have the required capability or you don't. I can see that in some environments (e.g., rsh across unmanaged collections of machines), it might be possible for someone to launch across a set of machines where some do and some don't have the required support. However, in all other cases, this will be homogeneous across the system. Given this analysis (and someone more familiar with the PML should feel free to confirm or correct it), it seems to me that this could be streamlined via one or more means: 1. at the most, we could have rank=0 add the PML module name to the modex, and other procs simply check it against their own and return an error if they differ. This accomplishes the identical functionality to what we have today, but with much less info in the modex. 2. we could eliminate this info from the modex altogether by requiring the user to specify the PML module if they want something other than the default OB1. In this case, there can be no confusion over what each proc is to use. The CM module will attempt to init the MTL - if it cannot do so, then the job will return the correct error and tell the user that CM/MTL support is unavailable. 3. we could again eliminate the info by not inserting it into the modex if (a) the default PML module is selected, or (b) the user specified the PML module to be used. In the first case, each proc can simply check to see if they picked the default - if not, then we can insert the info to indicate the difference. Thus, in the "standard" case, no info will be inserted. In the second case, we will already get an error if the specified PML module could not be used. Hence, the modex check provides no additional info or value. I understand the motivation to support automation. However, in this case, the automation actually doesn't seem to buy us very much, and it isn't coming "free". So perhaps some change in how this is done would be in order? Ralph ___ devel mailing list de...@open-mpi.org http://www.open-mpi.org/mailman/listinfo.cgi/devel
Re: [OMPI devel] ompi_ignore dr pml?
Until someone can work on it, sure, ompi_ignore DR sounds right. Unfortunately, IU may *need* to work on it this fall... hopefully we (I) will have a new student to help do the work. As for inclusion in 1.3, I don't think we care. On Mon, Jun 23, 2008 at 11:01 AM, Jeff Squyreswrote: > Should we .ompi_ignore dr? > > It's not complete and no one wants to support it. I'm thinking that we > shouldn't even include it in v1.3. > > Thoughts? > > -- > Jeff Squyres > Cisco Systems > > ___ > devel mailing list > de...@open-mpi.org > http://www.open-mpi.org/mailman/listinfo.cgi/devel -- Tim Mattox, Ph.D. - http://homepage.mac.com/tmattox/ tmat...@gmail.com || timat...@open-mpi.org I'm a bright... http://www.the-brights.net/
[OMPI devel] PML selection logic
Yo all I've been doing further research into the modex and came across something I don't fully understand. It seems we have each process insert into the modex the name of the PML module that it selected. Once the modex has exchanged that info, it then loops across all procs in the job to check their selection, and aborts if any proc picked a different PML module. All well and good...assuming that procs actually -can- choose different PML modules and hence create an "abort" scenario. However, if I look inside the PML's at their selection logic, I find that a proc can ONLY pick a module other than ob1 if: 1. the user specifies the module to use via -mca pml xyz or by using a module specific mca param to adjust its priority. In this case, since the mca param is propagated, ALL procs have no choice but to pick that same module, so that can't cause us to abort (we will have already returned an error and aborted if the specified module can't run). 2. the pml/cm module detects that an MTL module was selected, and that it is other than "psm". In this case, the CM module will be selected because its default priority is higher than that of OB1. In looking deeper into the MTL selection logic, it appears to me that you either have the required capability or you don't. I can see that in some environments (e.g., rsh across unmanaged collections of machines), it might be possible for someone to launch across a set of machines where some do and some don't have the required support. However, in all other cases, this will be homogeneous across the system. Given this analysis (and someone more familiar with the PML should feel free to confirm or correct it), it seems to me that this could be streamlined via one or more means: 1. at the most, we could have rank=0 add the PML module name to the modex, and other procs simply check it against their own and return an error if they differ. This accomplishes the identical functionality to what we have today, but with much less info in the modex. 2. we could eliminate this info from the modex altogether by requiring the user to specify the PML module if they want something other than the default OB1. In this case, there can be no confusion over what each proc is to use. The CM module will attempt to init the MTL - if it cannot do so, then the job will return the correct error and tell the user that CM/MTL support is unavailable. 3. we could again eliminate the info by not inserting it into the modex if (a) the default PML module is selected, or (b) the user specified the PML module to be used. In the first case, each proc can simply check to see if they picked the default - if not, then we can insert the info to indicate the difference. Thus, in the "standard" case, no info will be inserted. In the second case, we will already get an error if the specified PML module could not be used. Hence, the modex check provides no additional info or value. I understand the motivation to support automation. However, in this case, the automation actually doesn't seem to buy us very much, and it isn't coming "free". So perhaps some change in how this is done would be in order? Ralph
Re: [OMPI devel] multiple GigE interfaces...
On Wed, Jun 18, 2008 at 05:13:28PM -0700, Muhammad Atif wrote: > Hi again... I was on a break from Xensocket stuff This time some > general questions... Hi. > question. What if I have multiple Ethernet cards (say 5) on two of my > quad core machines. The IP addresses (and the subnets of course) are > Machine A Machine B > eth0 is y.y.1.a y.y.1.z > eth1 is y.y.4.by.y.4.y > eth2 is y.y.4.c ... > eth3 is y.y.4.d ... > > ... This sounds pretty weird. And I guess your netmasks don't allow to separate the NICs, do they? > from the FAQ's/Some emails in user lists it is clear that if I want > to run a job on multiple ethernets, I can use --mca btl_tcp_if_include > eth0,eth1. This You can, but you don't have to. If you don't specify something, OMPI will choose "something right". > will run the job on two of the subnets utilizing both the Ethernet > cards. Is it doing some sort of load balancing? or some round robin > mechanism? What part of code is responsible for this work? As far as I know, it's handled by OB1 (PML), which does striping across several BTL instances. So in other words, as long as both segments are equally fast, the load balancing should do fine. If they differ in performance, the OB1 doesn't find an optimal solution. If you're hitting this case, ask htor, he has an auto-tuning replacement, but that's not going to be part of OMPI. > eth1,eth2,eth3,eth4. Notice that all of these ethNs are on same subnet. > Even in the FAQ's (which mostly answers our lame questions) its not > entirely clear how communication will be done. Each process will have > tcp_num_btls equal to interfaces, but then what? Is it some sort of > load balancing or similar stuff which is not clear in tcpdump? I feel you could end up with communication stalls, the typical hang situation. One problem that might occur: the TCP component looks for remote addresses on the "same" network, so the component might be unable to decide whether your IP is on the same physical network or uses the wrong link. Then, you won't gain anything. Another problem: at least the Linux kernel (without tweaking) decides which interface and address to use for outgoing communication. If you have multiple subnets, then the kernel would go for the closest match between local and remote addresses, but in your case, it might be some kind of lottery. > related question is what if I want to run 8 process job (on 2x4 > cluster) and want to pin a process to an network interface. OpenMPI to > my understanding does not give any control of allocating IP to a > process (like MPICH) You could just say btl_if_include=ethX, thus giving you the right network interface. Obviously, this requires separate networks. > or is there some magical --mca thingie. I think only way to go is > adding routing tables... am i thinking in right direction? If yes, then > the performance of my boxes decrease when i trying to force the routing Routing should be fast, since it's done at kernel level. I cannot speak for Xen-based virtual interfaces. -- Cluster and Metacomputing Working Group Friedrich-Schiller-Universität Jena, Germany private: http://adi.thur.de
Re: [OMPI devel] BW benchmark hangs after r 18551
Hi, Seqf bug fixed in r18706. Best Regards Lenny. On Thu, Jun 19, 2008 at 5:37 PM, Lenny Verkhovsky < lenny.verkhov...@gmail.com> wrote: > Sorry, > I checked it without sm. > > pls ignore this mail. > > > > On Thu, Jun 19, 2008 at 4:32 PM, Lenny Verkhovsky < > lenny.verkhov...@gmail.com> wrote: > >> Hi, >> I found what caused the problem in both cases. >> >> --- ompi/mca/btl/sm/btl_sm.c(revision 18675) >> +++ ompi/mca/btl/sm/btl_sm.c(working copy) >> @@ -812,7 +812,7 @@ >> */ >> MCA_BTL_SM_FIFO_WRITE(endpoint, endpoint->my_smp_rank, >>endpoint->peer_smp_rank, frag->hdr, false, rc); >> -return (rc < 0 ? rc : 1); >> + return OMPI_SUCCESS; >> } >> I am just not sure if it's OK. >> >> Lenny. >> On Wed, Jun 18, 2008 at 3:21 PM, Lenny Verkhovsky < >> lenny.verkhov...@gmail.com> wrote: >> >>> Hi, >>> I am not sure if it related, >>> but I applied your patch ( r18667 ) to r 18656 ( one before NUMA ) >>> together with disabling sendi, >>> The result still the same ( hanging ). >>> >>> >>> >>> >>> On Tue, Jun 17, 2008 at 2:10 PM, George Bosilca>>> wrote: >>> Lenny, I guess you're running the latest version. If not, please update, Galen and myself corrected some bugs last week. If you're using the latest (and greatest) then ... well I imagine there is at least one bug left. There is a quick test you can do. In the btl_sm.c in the module structure at the beginning of the file, please replace the sendi function by NULL. If this fix the problem, then at least we know that it's a sm send immediate problem. Thanks, george. On Jun 17, 2008, at 7:54 AM, Lenny Verkhovsky wrote: Hi, George, > > I have a problem running BW benchmark on 100 rank cluster after r18551. > The BW is mpi_p that runs mpi_bandwidth with 100K between all pairs. > > > #mpirun -np 100 -hostfile hostfile_w ./mpi_p_18549 -t bw -s 10 > BW (100) (size min max avg) 10 576.734030 2001.882416 > 1062.698408 > #mpirun -np 100 -hostfile hostfile_w ./mpi_p_18551 -t bw -s 10 > mpirun: killing job... > ( it hangs even after 10 hours ). > > > It doesn't happen if I run --bynode or btl openib,self only. > > > Lenny. > >>> >> >