From: Andi K. <an...@fi...> - 2009-07-28 06:13:34
|
Fix a wrong Nehalem event Somehow event 0xa7 got the wrong name Still a 0.9.5 candidate Signed-off-by: Andi Kleen <ak...@li...> --- ChangeLog | 3 +++ events/i386/nehalem/events | 2 +- 2 files changed, 4 insertions(+), 1 deletion(-) Index: oprofile-0.9.5-rc2/events/i386/nehalem/events =================================================================== --- oprofile-0.9.5-rc2.orig/events/i386/nehalem/events +++ oprofile-0.9.5-rc2/events/i386/nehalem/events @@ -69,7 +69,7 @@ event:0x88 counters:0,1,2,3 um:br_inst_e event:0x89 counters:0,1,2,3 um:br_misp_exec minimum:6000 name:BR_MISP_EXEC : Counts the number of mispredicted conditional near branch instructions executed, but not necessarily retired. event:0xA2 counters:0,1,2,3 um:resource_stalls minimum:6000 name:RESOURCE_STALLS : Counts the number of Allocator resource related stalls. Includes register renaming buffer entries, memory buffer entries. In addition to resource related stalls, this event counts some other events. Includes stalls arising during branch misprediction recovery, such as if retirement of the mispredicted branch is delayed and stalls arising while store buffer is draining from synchronizing operations. event:0xA6 counters:0,1,2,3 um:one minimum:6000 name:MACRO_INSTS : Counts the number of instructions decoded that are macro-fused but not necessarily executed or retired. -event:0xA7 counters:0,1,2,3 um:one minimum:6000 name:ONE : Counts number of times a BACLEAR was forced by the Instruction Queue. The IQ is also responsible for providing conditional branch prediciton direction based on a static scheme and dynamic data provided by the L2 Branch Prediction Unit. If the conditional branch target is not found in the Target Array and the IQ predicts that the branch is taken, then the IQ will force the Branch Address Calculator to issue a BACLEAR. Each BACLEAR asserted by the BAC generates approximately an 8 cycle bubble in the instruction fetch pipeline. +event:0xA7 counters:0,1,2,3 um:one minimum:6000 name:BACLEAR_FORCE_IQ : Counts number of times a BACLEAR was forced by the Instruction Queue. The IQ is also responsible for providing conditional branch prediciton direction based on a static scheme and dynamic data provided by the L2 Branch Prediction Unit. If the conditional branch target is not found in the Target Array and the IQ predicts that the branch is taken, then the IQ will force the Branch Address Calculator to issue a BACLEAR. Each BACLEAR asserted by the BAC generates approximately an 8 cycle bubble in the instruction fetch pipeline. event:0xA8 counters:0,1,2,3 um:one minimum:6000 name:LSD : Counts the number of micro-ops delivered by loop stream detector event:0xAE counters:0,1,2,3 um:one minimum:6000 name:ITLB_FLUSH : Counts the number of ITLB flushes event:0xB0 counters:0,1,2,3 um:offcore_requests minimum:6000 name:OFFCORE_REQUESTS : Counts number of offcore data requests. Index: oprofile-0.9.5-rc2/ChangeLog =================================================================== --- oprofile-0.9.5-rc2.orig/ChangeLog +++ oprofile-0.9.5-rc2/ChangeLog @@ -1,3 +1,6 @@ +2009-07-28 Andi Kleen <ak...@li...> + * events/i386/nehalem/events: Fix name of event 0xa7. + 2009-07-20 Jeffrey Yasskin <jya...@go...> * libutil++/bfd_support.h(find_nearest_line): |
From: Maynard J. <may...@us...> - 2009-07-30 18:09:25
|
Andi Kleen wrote: > Fix a wrong Nehalem event > > Somehow event 0xa7 got the wrong name > > Still a 0.9.5 candidate > > Signed-off-by: Andi Kleen <ak...@li...> > Andi, I'm inclined to accept this patch and then release 0.9.5 GA, instead of going through another RC iteration. The risk seems negligible, wouldn't you agree? I would ask, though, if you could test the GA release on Nehalem ASAP just to be safe. Thanks. -Maynard > --- > ChangeLog | 3 +++ > events/i386/nehalem/events | 2 +- > 2 files changed, 4 insertions(+), 1 deletion(-) > > Index: oprofile-0.9.5-rc2/events/i386/nehalem/events > =================================================================== > --- oprofile-0.9.5-rc2.orig/events/i386/nehalem/events > +++ oprofile-0.9.5-rc2/events/i386/nehalem/events > @@ -69,7 +69,7 @@ event:0x88 counters:0,1,2,3 um:br_inst_e > event:0x89 counters:0,1,2,3 um:br_misp_exec minimum:6000 name:BR_MISP_EXEC : Counts the number of mispredicted conditional near branch instructions executed, but not necessarily retired. > event:0xA2 counters:0,1,2,3 um:resource_stalls minimum:6000 name:RESOURCE_STALLS : Counts the number of Allocator resource related stalls. Includes register renaming buffer entries, memory buffer entries. In addition to resource related stalls, this event counts some other events. Includes stalls arising during branch misprediction recovery, such as if retirement of the mispredicted branch is delayed and stalls arising while store buffer is draining from synchronizing operations. > event:0xA6 counters:0,1,2,3 um:one minimum:6000 name:MACRO_INSTS : Counts the number of instructions decoded that are macro-fused but not necessarily executed or retired. > -event:0xA7 counters:0,1,2,3 um:one minimum:6000 name:ONE : Counts number of times a BACLEAR was forced by the Instruction Queue. The IQ is also responsible for providing conditional branch prediciton direction based on a static scheme and dynamic data provided by the L2 Branch Prediction Unit. If the conditional branch target is not found in the Target Array and the IQ predicts that the branch is taken, then the IQ will force the Branch Address Calculator to issue a BACLEAR. Each BACLEAR asserted by the BAC generates approximately an 8 cycle bubble in the instruction fetch pipeline. > +event:0xA7 counters:0,1,2,3 um:one minimum:6000 name:BACLEAR_FORCE_IQ : Counts number of times a BACLEAR was forced by the Instruction Queue. The IQ is also responsible for providing conditional branch prediciton direction based on a static scheme and dynamic data provided by the L2 Branch Prediction Unit. If the conditional branch target is not found in the Target Array and the IQ predicts that the branch is taken, then the IQ will force the Branch Address Calculator to issue a BACLEAR. Each BACLEAR asserted by the BAC generates approximately an 8 cycle bubble in the instruction fetch pipeline. > event:0xA8 counters:0,1,2,3 um:one minimum:6000 name:LSD : Counts the number of micro-ops delivered by loop stream detector > event:0xAE counters:0,1,2,3 um:one minimum:6000 name:ITLB_FLUSH : Counts the number of ITLB flushes > event:0xB0 counters:0,1,2,3 um:offcore_requests minimum:6000 name:OFFCORE_REQUESTS : Counts number of offcore data requests. > Index: oprofile-0.9.5-rc2/ChangeLog > =================================================================== > --- oprofile-0.9.5-rc2.orig/ChangeLog > +++ oprofile-0.9.5-rc2/ChangeLog > @@ -1,3 +1,6 @@ > +2009-07-28 Andi Kleen <ak...@li...> > + * events/i386/nehalem/events: Fix name of event 0xa7. > + > 2009-07-20 Jeffrey Yasskin <jya...@go...> > > * libutil++/bfd_support.h(find_nearest_line): |
From: Andi K. <an...@fi...> - 2009-07-31 09:54:34
|
On Thu, Jul 30, 2009 at 01:09:26PM -0500, Maynard Johnson wrote: > Andi Kleen wrote: > > Fix a wrong Nehalem event > > > > Somehow event 0xa7 got the wrong name > > > > Still a 0.9.5 candidate > > > > Signed-off-by: Andi Kleen <ak...@li...> > > > Andi, > I'm inclined to accept this patch and then release 0.9.5 GA, instead of going through another RC iteration. The risk seems negligible, wouldn't you agree? I would ask, though, if you could test the GA release on Nehalem ASAP just to be safe. Ok. -Andi -- ak...@li... -- Speaking for myself only. |