Skip to contentSkip to conversation

SMPDesign: Remove duplicate item

PATCH In review Alan Huang 2 participants 5 messages Apr 11, 2023 – Apr 12, 2023

lore
Share & tools
lore

Alan Huang

author
#0 lore raw

PATCH 1/1 SMPDesign: Remove duplicate item

---
 SMPdesign/criteria.tex | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)
SMPdesign/criteria.tex +1 -6
index d3d84506..bc4d3a77 100644
@@ -145,7 +145,7 @@ parallel program.
 	The larger the gap between the number of CPUs
 	and the actual speedup, the less efficiently the
 	CPUs will be used.
-	Similarly, the greater the desired efficiency, the smaller
+	Similarly, the greater the desired efficiency, the bigger
 	the achievable speedup.
 \item	If the available synchronization primitives have
 	high overhead compared to the critical sections
@@ -157,11 +157,6 @@ parallel program.
 	using asymmetric primitives
 	(see \cref{chp:Deferred Processing}),
 	or by using a coarse-grained design such as \IXh{code}{locking}.
-\item	If the critical sections have high overhead compared
-	to the primitives guarding them, the best way
-	to improve speedup is to increase parallelism
-	by moving to reader/writer locking, \IXh{data}{locking}, asymmetric,
-	or data ownership.
 \item	If the critical sections have high overhead compared
 	to the primitives guarding them and the data structure
 	being guarded is read much more often than modified,
Signature · 1 line
2.34.1
#1 lore raw

Hi Alan,

Quoted 10 lines · Alan Huang
---
 SMPdesign/criteria.tex | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)
SMPdesign/criteria.tex +0 -0
index d3d84506..bc4d3a77 100644

SMPdesign/criteria.tex parallel program. · line 149

Quoted 6 lines
SMPdesign/criteria.tex +1 -1
@@ -145,7 +145,7 @@ parallel program.
 	The larger the gap between the number of CPUs
 	and the actual speedup, the less efficiently the
 	CPUs will be used.
-	Similarly, the greater the desired efficiency, the smaller
+	Similarly, the greater the desired efficiency, the bigger
 	the achievable speedup.

I might not be fully woken up, but this change doesn't make sense to me.

SMPdesign/criteria.tex parallel program. · line 164

Quoted 11 lines
SMPdesign/criteria.tex +0 -5
 \item	If the available synchronization primitives have
 	high overhead compared to the critical sections
@@ -157,11 +157,6 @@ parallel program.
 	using asymmetric primitives
 	(see \cref{chp:Deferred Processing}),
 	or by using a coarse-grained design such as \IXh{code}{locking}.
-\item	If the critical sections have high overhead compared
-	to the primitives guarding them, the best way
-	to improve speedup is to increase parallelism
-	by moving to reader/writer locking, \IXh{data}{locking}, asymmetric,
-	or data ownership.

You mean, this and the next item says the same thing? I don't think so.

Item 4:

If the critical sections have high overhead compared to the primitives guarding them, the best way to improve
^^^^^^^
speedup is to increase parallelism by moving to reader/writer ^^^^^^^
locking, data locking, asymmetric, or data ownership.
^^^^^^^^^^^^^ ^^^^^^^^^^^^^^

Item 5:

If the critical sections have high overhead compared to the primitives guarding them and the data structure being
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
guarded is read much more often than modified, the best way ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
to increase parallelism is to move to reader/writer locking or asymmetric primitives.

Are you sure they are duplicates ???

Thanks, Akira

Quoted 3 lines
SMPdesign/criteria.tex +0 -0
 \item	If the critical sections have high overhead compared
 	to the primitives guarding them and the data structure
 	being guarded is read much more often than modified,
author
#2 lore raw
Quoted 6 lines
Quoted 3 lines
SMPdesign/criteria.tex +1 -1
-	Similarly, the greater the desired efficiency, the smaller
+	Similarly, the greater the desired efficiency, the bigger

the achievable speedup.

I might not be fully woken up, but this change doesn't make sense to me.

The original sentence means (to me) If we get greater efficiency use of CPUs, we get smaller speedup.

If I misunderstood, please correct me.

Quoted 12 lines

If the critical sections have high overhead compared to the primitives guarding them, the best way to improve

                                                 ^^^^^^^
   speedup is to increase parallelism by moving to reader/writer
   ^^^^^^^
   locking, data locking, asymmetric, or data ownership.
            ^^^^^^^^^^^^^                ^^^^^^^^^^^^^^

Item 5:

If the critical sections have high overhead compared to the primitives guarding them and the data structure being

Quoted 7 lines
                                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^

guarded is read much more often than modified, the best way ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
to increase parallelism is to move to reader/writer locking or asymmetric primitives.

Are you sure they are duplicates ???

If the critical sections have high overhead compared to the primitives guarding them,
It means that the overhead of primitives is relative small, but the aim of data locking and data ownership is to reduce the overhead of primitives.

Thanks,
Alan

Quoted 67 lines

2023年4月12日 上午8:16,Akira Yokosawa <[email protected]> 写道:

Hi Alan,

Quoted 10 lines · Alan Huang

---
SMPdesign/criteria.tex | 7 +------

1 file changed, 1 insertion(+), 6 deletions(-)
SMPdesign/criteria.tex +0 -0
index d3d84506..bc4d3a77 100644

SMPdesign/criteria.tex parallel program. · line 149

Quoted 6 lines
SMPdesign/criteria.tex +1 -1
@@ -145,7 +145,7 @@ parallel program.
 	The larger the gap between the number of CPUs
 	and the actual speedup, the less efficiently the
 	CPUs will be used.
-	Similarly, the greater the desired efficiency, the smaller
+	Similarly, the greater the desired efficiency, the bigger
 	the achievable speedup.

I might not be fully woken up, but this change doesn't make sense to me.

SMPdesign/criteria.tex parallel program. · line 164

Quoted 11 lines

\item If the available synchronization primitives have
high overhead compared to the critical sections

SMPdesign/criteria.tex +0 -5
@@ -157,11 +157,6 @@ parallel program.
 	using asymmetric primitives
 	(see \cref{chp:Deferred Processing}),
 	or by using a coarse-grained design such as \IXh{code}{locking}.
-\item	If the critical sections have high overhead compared
-	to the primitives guarding them, the best way
-	to improve speedup is to increase parallelism
-	by moving to reader/writer locking, \IXh{data}{locking}, asymmetric,
-	or data ownership.

You mean, this and the next item says the same thing? I don't think so.

Item 4:

If the critical sections have high overhead compared to the primitives guarding them, the best way to improve

                                                 ^^^^^^^
   speedup is to increase parallelism by moving to reader/writer
   ^^^^^^^
   locking, data locking, asymmetric, or data ownership.
            ^^^^^^^^^^^^^                ^^^^^^^^^^^^^^

Item 5:

If the critical sections have high overhead compared to the primitives guarding them and the data structure being

                                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^

guarded is read much more often than modified, the best way ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
to increase parallelism is to move to reader/writer locking or asymmetric primitives.

Are you sure they are duplicates ???

Thanks, Akira

Quoted 3 lines

\item If the critical sections have high overhead compared
to the primitives guarding them and the data structure being guarded is read much more often than modified,

#3 lore raw
Quoted 11 lines · Alan Huang
Quoted 6 lines
Quoted 3 lines
SMPdesign/criteria.tex +1 -1
-	Similarly, the greater the desired efficiency, the smaller
+	Similarly, the greater the desired efficiency, the bigger

the achievable speedup.

I might not be fully woken up, but this change doesn't make sense to me.

The original sentence means (to me) If we get greater efficiency use of CPUs, we get smaller speedup.

If I misunderstood, please correct me.

My version of interpretation of the sentence:

Similarly, if our goal is greater efficiency, we might end up in a smaller speedup in the end.

One extreme approach for the greatest efficiency might be not to parallelize at all, to use a single CPU, and to get a negative speedup (i.e., speed down).

Quoted 19 lines
Quoted 19 lines

If the critical sections have high overhead compared to the primitives guarding them, the best way to improve

                                                 ^^^^^^^
   speedup is to increase parallelism by moving to reader/writer
   ^^^^^^^
   locking, data locking, asymmetric, or data ownership.
            ^^^^^^^^^^^^^                ^^^^^^^^^^^^^^

Item 5:

If the critical sections have high overhead compared to the primitives guarding them and the data structure being

                                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^

guarded is read much more often than modified, the best way ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
to increase parallelism is to move to reader/writer locking or asymmetric primitives.

Quoted 6 lines
Quoted 1 line

Are you sure they are duplicates ???

If the critical sections have high overhead compared to the primitives guarding them,
It means that the overhead of primitives is relative small, but the aim of data locking and data ownership is to reduce the overhead of primitives.

Well, I think Paul's motivation of special casing read-mostly data comes from his expertise in RCU.

I'd like to respect his motivation.

Anyway, this list is more of a guideline of approaches to parallel-programming design considerations.
Why do you want it to be so precise ???

Thanks, Akira

Quoted 4 lines

Thanks,
Alan

[...]

author
#4 lore raw
Quoted 3 lines

One extreme approach for the greatest efficiency might be not to parallelize at all, to use a single CPU, and to get a negative speedup (i.e., speed down).

But we can also get the greatest efficiency using data ownership under the maximum degree of parallelism.

Quoted 2 lines

Anyway, this list is more of a guideline of approaches to parallel-programming design considerations.

Got it.

Thanks,
Alan

Quoted 66 lines

2023年4月12日 上午10:51,Akira Yokosawa <[email protected]> 写道:

Quoted 11 lines · Alan Huang
Quoted 6 lines
Quoted 3 lines
SMPdesign/criteria.tex +1 -1
-	Similarly, the greater the desired efficiency, the smaller
+	Similarly, the greater the desired efficiency, the bigger

the achievable speedup.

I might not be fully woken up, but this change doesn't make sense to me.

The original sentence means (to me) If we get greater efficiency use of CPUs, we get smaller speedup.

If I misunderstood, please correct me.

My version of interpretation of the sentence:

Similarly, if our goal is greater efficiency, we might end up in a smaller speedup in the end.

One extreme approach for the greatest efficiency might be not to parallelize at all, to use a single CPU, and to get a negative speedup (i.e., speed down).

Quoted 19 lines
Quoted 19 lines

If the critical sections have high overhead compared to the primitives guarding them, the best way to improve

                                                ^^^^^^^
  speedup is to increase parallelism by moving to reader/writer
  ^^^^^^^
  locking, data locking, asymmetric, or data ownership.
           ^^^^^^^^^^^^^                ^^^^^^^^^^^^^^

Item 5:

If the critical sections have high overhead compared to the primitives guarding them and the data structure being

                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^

guarded is read much more often than modified, the best way ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
to increase parallelism is to move to reader/writer locking or asymmetric primitives.

Quoted 6 lines
Quoted 1 line

Are you sure they are duplicates ???

If the critical sections have high overhead compared to the primitives guarding them,
It means that the overhead of primitives is relative small, but the aim of data locking and data ownership is to reduce the overhead of primitives.

Well, I think Paul's motivation of special casing read-mostly data comes from his expertise in RCU.

I'd like to respect his motivation.

Anyway, this list is more of a guideline of approaches to parallel-programming design considerations.
Why do you want it to be so precise ???

Thanks, Akira

Quoted 4 lines

Thanks,
Alan

[...]