Index: doc/theses/thierry_delisle_PhD/thesis/text/eval_micro.tex =================================================================== --- doc/theses/thierry_delisle_PhD/thesis/text/eval_micro.tex (revision 0fec6c16191867e177606d305a5c86998e83b934) +++ doc/theses/thierry_delisle_PhD/thesis/text/eval_micro.tex (revision 1fcbce7cbae20080fd7f928bf78187ddccb6d794) @@ -42,5 +42,5 @@ Each experiment is run 15 times varying the number of processors depending on the two different computers. All experiments gather throughput data and secondary data for scalability or latency. -The data is graphed using a solid, a dashed and a dotted line, representing the median, maximum and minimum result respectively, where the minimum/maximum lines are referred to as the \emph{extremes}.\footnote{ +The data is graphed using a solid, a dashed, and a dotted line, representing the median, maximum and minimum result respectively, where the minimum/maximum lines are referred to as the \emph{extremes}.\footnote{ An alternative display is to use error bars with min/max as the bottom/top for the bar. However, this approach is not truly an error bar around a mean value and I felt the connected lines are easier to read.} @@ -102,6 +102,5 @@ \label{fig:cycle:code} %\end{figure}ll have a physical key so it's not urgent. - - +\bigskip %\begin{figure} \subfloat[][Throughput, 100 cycles per \proc]{ @@ -130,5 +129,5 @@ \label{fig:cycle:jax:low:ns} } - \caption[Cycle Benchmark on Intel]{Cycle Benchmark on Intel\smallskip\newline Throughput and scalability as a function of \proc count, 5 \ats per cycle, and different cycle count. + \caption[Cycle Benchmark on Intel]{Cycle Benchmark on Intel\smallskip\newline Throughput and scalability as a function of \proc count, 5 \ats per cycle, and different cycle counts. For throughput, higher is better, for scalability, lower is better. Each series represent 15 independent runs, the dashed lines are maximums of each series while the solid lines are the median and the dotted lines are the minimums.} @@ -170,6 +169,6 @@ \subsection{Results} -Figures~\ref{fig:cycle:jax} and \ref{fig:cycle:nasus} show the results for the cycle experiment. -Looking at the left column on Intel first, Figures~\ref{fig:cycle:jax:ops} and \ref{fig:cycle:jax:ns}, which shows the results for many \ats, in this case 100 cycles of 5 \ats for each \proc. +Figures~\ref{fig:cycle:jax} and \ref{fig:cycle:nasus} show the results for the cycle experiment on Intel and AMD, respectively. +Looking at the left column on Intel, Figures~\ref{fig:cycle:jax:ops} and \ref{fig:cycle:jax:ns} show the results for 100 cycles of 5 \ats for each \proc. \CFA, Go and Tokio all obtain effectively the same throughput performance. Libfibre is slightly behind in this case but still scales decently. @@ -178,19 +177,19 @@ As expected, this pattern repeats again between \proc count 72 and 96. -Looking next at the right column on Intel, Figures~\ref{fig:cycle:jax:low:ops} and \ref{fig:cycle:jax:low:ns}, which shows the results for few threads, in this case 1 cycle of 5 \ats for each \proc. +Looking next at the right column on Intel, Figures~\ref{fig:cycle:jax:low:ops} and \ref{fig:cycle:jax:low:ns} show the results for 1 cycle of 5 \ats for each \proc. \CFA and Tokio obtain very similar results overall, but Tokio shows more variations in the results. -Go achieves slightly better performance than \CFA and Tokio, but all three display significantly workst performance compared to the left column. +Go achieves slightly better performance than \CFA and Tokio, but all three display significantly worst performance compared to the left column. This decrease in performance is likely due to the additional overhead of the idle-sleep mechanism. This can either be the result of \procs actually running out of work, or simply additional overhead from tracking whether or not there is work available. -Indeed, unlike the left column, it is likely that the ready-queue will be transiently empty, which likely triggers additional synchronization steps. +Indeed, unlike the left column, it is likely that the ready-queue is transiently empty, which likely triggers additional synchronization steps. Interestingly, libfibre achieves better performance with 1 cycle. -Looking now at the results for the AMD architecture, Figure~\ref{fig:cycle:nasus}, the results show a story that is overall similar to the results on the Intel, with close to double the performance overall but with slightly increased variation and some differences in the details. -Note that the maximum of the Y-axis on Intel and AMD differ significantly. -Looking at the left column first, Figures~\ref{fig:cycle:nasus:ops} and \ref{fig:cycle:nasus:ns}, unlike Intel, on AMD all 4 runtimes achieve very similar throughput and scalability. +Looking now at the results for the AMD architecture, Figure~\ref{fig:cycle:nasus}, the results are overall similar to the Intel results, but with close to double the performance, slightly increased variation, and some differences in the details. +Note the maximum of the Y-axis on Intel and AMD differ significantly. +Looking at the left column on AMD, Figures~\ref{fig:cycle:nasus:ops} and \ref{fig:cycle:nasus:ns} all 4 runtimes achieve very similar throughput and scalability. However, as the number of \procs grows higher, the results on AMD show notably more variability than on Intel. -The different performance improvements and plateaus are due to cache topology and appear at the expected \proc counts of 64, 128 and 192, for the same reasons as on Intel. -Looking next at the right column, Figures~\ref{fig:cycle:nasus:low:ops} and \ref{fig:cycle:nasus:low:ns}, Tokio and Go have the same throughput performance, while \CFA is slightly slower. -This is different than on Intel, where Tokio behaved like \CFA rather than behaving like Go. +The different performance improvements and plateaus are due to cache topology and appear at the expected: \proc counts of 64, 128 and 192, for the same reasons as on Intel. +Looking next at the right column on AMD, Figures~\ref{fig:cycle:nasus:low:ops} and \ref{fig:cycle:nasus:low:ns}, Tokio and Go have the same throughput performance, while \CFA is slightly slower. +This result is different than on Intel, where Tokio behaved like \CFA rather than behaving like Go. Again, the same performance increase for libfibre is visible when running fewer \ats. Note, I did not investigate the libfibre performance boost for 1 cycle in this experiment. @@ -203,5 +202,5 @@ \section{Yield} -For completion, the classic yield benchmark is included. +For completeness, the classic yield benchmark is included. Here, the throughput is dominated by the mechanism used to handle the @yield@ function. Figure~\ref{fig:yield:code} shows pseudo code for this benchmark, where the cycle @wait/next.wake@ is replaced by @yield@. @@ -259,7 +258,7 @@ \subsection{Results} -Figures~\ref{fig:yield:jax} and \ref{fig:yield:nasus} show the results for the yield experiment. -Looking at the left column on Intel first, Figures~\ref{fig:yield:jax:ops} and \ref{fig:yield:jax:ns}, which shows the results for many \ats, in this case 100 \ats for each \proc. -Note, the Y-axis on the graph is twice as large as the Intel cycle-graph. +Figures~\ref{fig:yield:jax} and \ref{fig:yield:nasus} show the results for the yield experiment on Intel and AMD, respectively. +Looking at the left column on Intel, Figures~\ref{fig:yield:jax:ops} and \ref{fig:yield:jax:ns} show the results for 100 \ats for each \proc. +Note, the Y-axis on this graph is twice as large as the Intel cycle-graph. A visual glance between the left columns of the cycle and yield graphs confirms my claim that the yield benchmark is unreliable. \CFA has no special handling for @yield@, but this experiment requires less synchronization than the @cycle@ experiment. @@ -276,7 +275,7 @@ This lack of communication is probably why the plateaus due to topology are not present. -Lookking next at the right column on Intel, Figures~\ref{fig:yield:jax:low:ops} and \ref{fig:yield:jax:low:ns}, which shows the results for few threads, in this case 1 \at for each \proc. +Looking next at the right column on Intel, Figures~\ref{fig:yield:jax:low:ops} and \ref{fig:yield:jax:low:ns} show the results for 1 \at for each \proc. As for @cycle@, \CFA's cost of idle sleep comes into play in a very significant way in Figure~\ref{fig:yield:jax:low:ns}, where the scaling is not flat. -This is to be expected since fewet \ats means \procs are more likely to run out of work. +This result is to be expected since fewer \ats means \procs are more likely to run out of work. On the other hand, when only running 1 \at per \proc, libfibre optimizes further, and forgoes the context-switch entirely. This results in libfibre outperforming other runtimes even more, achieving 8 times more throughput than for @cycle@. @@ -320,9 +319,9 @@ Looking at the left column first, Figures~\ref{fig:yield:nasus:ops} and \ref{fig:yield:nasus:ns}, \CFA achieves very similar throughput and scaling. Libfibre still outpaces all other runtimes, but it encounter a performance hit at 64 \procs. -This suggest some amount of communication between the \procs that the Intel machine was able to mask where the AMD is not once hyperthreading is needed. -Go and Tokio still display the same performance collapse than on Intel. -Looking next at the right column, Figures~\ref{fig:yield:nasus:low:ops} and \ref{fig:yield:nasus:low:ns}, all runtime effectively behave the same as they did on the Intel machine. -At high \ats count the only difference was Libfibre's scaling and this difference disappears on the right column. -This suggest that whatever communication benchmark it encountered on the left is completely circumvented on the right. +This anomaly suggest some amount of communication between the \procs that the Intel machine is able to mask where the AMD is not once hyperthreading is needed. +Go and Tokio still display the same performance collapse as on Intel. +Looking next at the right column on AMD, Figures~\ref{fig:yield:nasus:low:ops} and \ref{fig:yield:nasus:low:ns}, all runtime systems effectively behave the same as they did on the Intel machine. +At the high \ats count, the only difference is Libfibre's scaling and this difference disappears on the right column. +This behaviour suggest whatever communication issue it encountered on the left is completely circumvented on the right. It is difficult to draw conclusions for this benchmark when runtime system treat @yield@ so differently. @@ -335,5 +334,5 @@ In these benchmarks, \ats can be easily partitioned over the different \procs upfront and none of the \ats communicate with each other. -The Churn benchmark represents more chaotic executions, where there is more communication among \ats but no relationship between the last \proc on which a \at ran and blocked and the \proc that subsequently unblocks it. +The Churn benchmark represents more chaotic executions, where there is more communication among \ats but no relationship between the last \proc on which a \at ran and blocked, and the \proc that subsequently unblocks it. With processor-specific ready-queues, when a \at is unblocked by a different \proc that means the unblocking \proc must either ``steal'' the \at from another processor or find it on a remote queue. This dequeuing results in either contention on the remote queue and/or \glspl{rmr} on the \at data structure. @@ -342,5 +341,5 @@ This benchmark uses a fixed-size array of counting semaphores. -Each \at picks a random semaphore, @V@s it to unblock any waiting \at, and then @P@s (maybe blocks) the \ats on the semaphore. +Each \at picks a random semaphore, @V@s it to unblock any waiting \at, and then @P@s (maybe blocks) the \at on the semaphore. This creates a flow where \ats push each other out of the semaphores before being pushed out themselves. For this benchmark to work, the number of \ats must be equal or greater than the number of semaphores plus the number of \procs; @@ -393,5 +392,5 @@ \label{fig:churn:jax:low:ns} } - \caption[Churn Benchmark on Intel]{\centering Churn Benchmark on Intel\smallskip\newline Throughput and scalability of the Churn on the benchmark on the Intel machine. + \caption[Churn Benchmark on Intel]{Churn Benchmark on Intel\smallskip\newline Throughput and scalability of the Churn on the benchmark on the Intel machine. For throughput, higher is better, for scalability, lower is better. Each series represent 15 independent runs, the dashed lines are maximums of each series while the solid lines are the median and the dotted lines are the minimums.} @@ -401,30 +400,30 @@ \subsection{Results} -Figures~\ref{fig:churn:jax} and Figure~\ref{fig:churn:nasus} show the results for the churn experiment. -Looking at the left column on Intel first, Figures~\ref{fig:churn:jax:ops} and \ref{fig:churn:jax:ns}, which shows the results for many \ats, in this case 100 \ats for each \proc, all runtime obtain fairly similar throughput for most \proc counts. +Figures~\ref{fig:churn:jax} and Figure~\ref{fig:churn:nasus} show the results for the churn experiment on Intel and AMD, respectively. +Looking at the left column on Intel, Figures~\ref{fig:churn:jax:ops} and \ref{fig:churn:jax:ns} show the results for 100 \ats for each \proc have, and all runtimes obtain fairly similar throughput for most \proc counts. \CFA does very well on a single \proc but quickly loses its advantage over the other runtimes. -As expected it scales decently up to 48 \procs and then basically plateaus. -Tokio achieves very similar performance to \CFA until 48 \procs, after which it takes a significant hit but does keep scaling somewhat. -Libfibre obtains effectively the same results as Tokio with slightly less scaling, \ie the scaling curve is the same but with slightly higher values. -Finally Go gets the most peculiar results, scaling worst than other runtimes until 48 \procs. +As expected, it scales decently up to 48 \procs, drops from 48 to 72 \procs, and then plateaus. +Tokio achieves very similar performance to \CFA, with the starting boost, scaling decently until 48 \procs, drops from 48 to 72 \procs, and starts increasing again to 192 \procs. +Libfibre obtains effectively the same results as Tokio with slightly less scaling, \ie the scaling curve is the same but with slightly lower values. +Finally, Go gets the most peculiar results, scaling worst than other runtimes until 48 \procs. At 72 \procs, the results of the Go runtime vary significantly, sometimes scaling sometimes plateauing. -However, beyond this point Go keeps this level of variation but does not scale in any of the runs. +However, beyond this point Go keeps this level of variation but does not scale further in any of the runs. Throughput and scalability is notably worst for all runtimes than the previous benchmarks since there is inherently more communication between processors. -Indeed, none of the runtime reach 40 million operations per second while in the cycle benchmark all but libfibre reached 400 million operations per second. -Figures~\ref{fig:churn:jax:ns} and \ref{fig:churn:jax:low:ns} show that for all \proc count, all runtime produce poor scaling. +Indeed, none of the runtimes reach 40 million operations per second while in the cycle benchmark all but libfibre reached 400 million operations per second. +Figures~\ref{fig:churn:jax:ns} and \ref{fig:churn:jax:low:ns} show that for all \proc count, all runtimes produce poor scaling. However, once the number of \glspl{hthrd} goes beyond a single socket, at 48 \procs, scaling goes from bad to worst and performance completely ceases to improve. At this point, the benchmark is dominated by inter-socket communication costs for all runtimes. An interesting aspect to note here is that the runtimes differ in how they handle this situation. -Indeed, when a \proc unparks a \at that was last run on a different \proc, the \at could be appended to the ready-queue local \proc or to the ready-queue of the remote \proc, which previously ran the \at. -\CFA, Tokio and Go all use the approach of unparking to the local \proc while Libfibre unparks to the remote \proc. -In this particular benchmark, the inherent chaos of the benchmark in addition to small memory footprint means neither approach wins over the other. - -Looking next at the right column on Intel, Figures~\ref{fig:churn:jax:low:ops} and \ref{fig:churn:jax:low:ns}, which shows the results for few threads, in this case 1 \at for each \proc, many of the differences between the runtime disappear. +Indeed, when a \proc unparks a \at that was last run on a different \proc, the \at could be appended to the ready queue of the local \proc or to the ready queue of the remote \proc, which previously ran the \at. +\CFA, Tokio and Go all use the approach of unparking to the local \proc, while Libfibre unparks to the remote \proc. +In this particular benchmark, the inherent chaos of the benchmark, in addition to small memory footprint, means neither approach wins over the other. + +Looking next at the right column on Intel, Figures~\ref{fig:churn:jax:low:ops} and \ref{fig:churn:jax:low:ns} show the results for 1 \at for each \proc, and many of the differences between the runtimes disappear. \CFA outperforms other runtimes by a minuscule margin. Libfibre follows very closely behind with basically the same performance and scaling. -Tokio maintains effectively the same curve shapes as it did with many threads, but it incurs extra costs for all \proc count. -As a result it is slightly outperformed by \CFA and libfibre. +Tokio maintains effectively the same curve shapes as \CFA and libfibre, but it incurs extra costs for all \proc counts. +% As a result it is slightly outperformed by \CFA and libfibre. While Go maintains overall similar results to the others, it again encounters significant variation at high \proc counts. Inexplicably resulting in super-linear scaling for some runs, \ie the scalability curves displays a negative slope. @@ -459,5 +458,5 @@ \label{fig:churn:nasus:low:ns} } - \caption[Churn Benchmark on AMD]{\centering Churn Benchmark on AMD\smallskip\newline Throughput and scalability of the Churn on the benchmark on the AMD machine. + \caption[Churn Benchmark on AMD]{Churn Benchmark on AMD\smallskip\newline Throughput and scalability of the Churn on the benchmark on the AMD machine. For throughput, higher is better, for scalability, lower is better. Each series represent 15 independent runs, the dashed lines are maximums of each series while the solid lines are the median and the dotted lines are the minimums.} @@ -468,21 +467,20 @@ Looking now at the results for the AMD architecture, Figure~\ref{fig:churn:nasus}, the results show a somewhat different story. Looking at the left column first, Figures~\ref{fig:churn:nasus:ops} and \ref{fig:churn:nasus:ns}, \CFA, Libfibre and Tokio all produce decent scalability. -\CFA suffers particular from a larger variations at higher \proc counts, but almost all run still outperform the other runtimes. +\CFA suffers particular from a larger variations at higher \proc counts, but largely outperforms the other runtimes. Go still produces intriguing results in this case and even more intriguingly, the results have fairly low variation. -One possible explanation for this difference is that since Go has very few available concurrent primitives, a channel was used instead of a semaphore. -On paper a semaphore can be replaced by a channel and with zero-sized objects passed along equivalent performance could be expected. -However, in practice there can be implementation difference between the two. -This is especially true if the semaphore count can get somewhat high. -Note that this replacement is also made in the cycle benchmark, however in that context it did not seem to have a notable impact. - -As second possible explanation is that Go may sometimes use the heap when allocating variables based on the result of escape analysis of the code. -It is possible that variables that should be placed on the stack are placed on the heap. -This could cause extra pointer chasing in the benchmark, heightening locality effects. +One possible explanation for Go's difference is that it has very few available concurrent primitives, so a channel is substituted for a semaphore. +On paper a semaphore can be replaced by a channel, and with zero-sized objects passed through the channel, equivalent performance could be expected. +However, in practice, there are implementation difference between the two, \eg if the semaphore count can get somewhat high so object accumulate in the channel. +Note that this substitution is also made in the cycle benchmark; +however, in that context, it did not have a notable impact. + +As second possible explanation is that Go may use the heap when allocating variables based on the result of escape analysis of the code. +It is possible for variables that could be placed on the stack to instead be placed on the heap. +This placement could cause extra pointer chasing in the benchmark, heightening locality effects. Depending on how the heap is structure, this could also lead to false sharing. - -I did not further investigate what causes these unusual results. - -Looking next at the right column, Figures~\ref{fig:churn:nasus:low:ops} and \ref{fig:churn:nasus:low:ns}, like for Intel all runtime obtain overall similar throughput between the left and right column. +I did not investigate what causes these unusual results. + +Looking next at the right column, Figures~\ref{fig:churn:nasus:low:ops} and \ref{fig:churn:nasus:low:ns}, as for Intel, all runtimes obtain overall similar throughput between the left and right column. \CFA, Libfibre and Tokio all have very close results. Go still suffers from poor scalability but is now unusual in a different way. @@ -490,9 +488,27 @@ Up to 32 \procs, after which the other runtime manage to outscale Go. -The objective of this benchmark is to demonstrate that unparking \ats from remote \procs do not cause too much contention on the local queues. -Indeed, the fact most runtimes achieve some scaling between various \proc count demonstrate that migrations do not need to be serialized. -Again these result demonstrate \CFA achieves satisfactory performance. +In conclusion, the objective of this benchmark is to demonstrate that unparking \ats from remote \procs does not cause too much contention on the local queues. +Indeed, the fact that most runtimes achieve some scaling between various \proc count demonstrate migrations do not need to be serialized. +Again these result demonstrate \CFA achieves satisfactory performance with respect to the other runtimes. \section{Locality} + +As mentioned in the churn benchmark, when unparking a \at, it is possible to either unpark to the local or remote ready-queue.\footnote{ +It is also possible to unpark to a third unrelated ready-queue, but without additional knowledge about the situation, it is likely to degrade performance.} +The locality experiment includes two variations of the churn benchmark, where a data array is added. +In both variations, before @V@ing the semaphore, each \at increments random cells inside the data array by calling a @work@ function. +In the noshare variation, the array is not passed on and each thread continuously accesses its private array. +In the share variation, the array is passed to another thread via the semaphore's shadow-queue (each blocking thread can save a word of user data in its blocking node), transferring ownership of the array to the woken thread. +Figure~\ref{fig:locality:code} shows pseudo code for this benchmark. + +The objective here is to highlight the different decision made by the runtime when unparking. +Since each thread unparks a random semaphore, it means that it is unlikely that a \at is unparked from the last \proc it ran on. +In the noshare variation, unparking the \at on the local \proc is an appropriate choice since the data was last modified on that \proc. +In the shared variation, unparking the \at on a remote \proc is an appropriate choice. +\todo{PAB: I changed these sentences around.} + +The expectation for this benchmark is to see a performance inversion, where runtimes fare notably better in the variation which matches their unparking policy. +This decision should lead to \CFA, Go and Tokio achieving better performance in the share variation while libfibre achieves better performance in noshare. +Indeed, \CFA, Go and Tokio have the default policy of unparking \ats on the local \proc, where as libfibre has the default policy of unparking \ats wherever they last ran. \begin{figure} @@ -538,9 +554,9 @@ \end{lrbox} -\subfloat[Thread$_1$]{\label{f:CFibonacci}\usebox\myboxA} +\subfloat[Noshare]{\label{fig:locality:code:T1}\usebox\myboxA} \hspace{3pt} \vrule \hspace{3pt} -\subfloat[Thread$_2$]{\label{f:CFAFibonacciGen}\usebox\myboxB} +\subfloat[Share]{\label{fig:locality:code:T1}\usebox\myboxB} \caption[Locality Benchmark : Pseudo Code]{Locality Benchmark : Pseudo Code} @@ -548,21 +564,15 @@ \end{figure} -As mentioned in the churn benchmark, when unparking a \at, it is possible to either unpark to the local or remote ready-queue. -\footnote{It is also possible to unpark to a third unrelated ready-queue, but without additional knowledge about the situation, there is little to suggest this would not degrade performance.} -The locality experiment includes two variations of the churn benchmark, where an array of data is added. -In both variations, before @V@ing the semaphore, each \at increment random cells inside the array. -The @share@ variation then passes the array to the shadow-queue of the semaphore, transferring ownership of the array to the woken thread. -In the @noshare@ variation the array is not passed on and each thread continuously accesses its private array. - -The objective here is to highlight the different decision made by the runtime when unparking. -Since each thread unparks a random semaphore, it means that it is unlikely that a \at will be unparked from the last \proc it ran on. -In the @share@ version, this means that unparking the \at on the local \proc is appropriate since the data was last modified on that \proc. -In the @noshare@ version, the unparking the \at on the remote \proc is the appropriate approach. - -The expectation for this benchmark is to see a performance inversion, where runtimes will fare notably better in the variation which matches their unparking policy. -This should lead to \CFA, Go and Tokio achieving better performance in @share@ while libfibre achieves better performance in @noshare@. -Indeed, \CFA, Go and Tokio have the default policy of unpark \ats on the local \proc, where as libfibre has the default policy of unparks \ats wherever they last ran. - \subsection{Results} + +Figures~\ref{fig:locality:jax} and \ref{fig:locality:nasus} show the results for the locality experiment on Intel and AMD, respectively. +In both cases, the graphs on the left column show the results for the share variation and the graphs on the right column show the results for the noshare. +Looking at the left column on Intel, Figures~\ref{fig:locality:jax:share:ops} and \ref{fig:locality:jax:share:ns} show the results for the share variation. +\CFA and Tokio slightly outperform libfibre, as expected, based on their \ats placement approach. +\CFA and Tokio both unpark locally and do not suffer cache misses on the transferred array. +Libfibre on the other hand unparks remotely, and as such the unparked \at is likely to miss on the shared data. +Go trails behind in this experiment, presumably for the same reasons that were observable in the churn benchmark. +Otherwise, the results are similar to the churn benchmark, with lower throughput due to the array processing. +As for most previous results, all runtimes suffer a performance hit after 48 \proc, which is the socket boundary, and climb again from 96 to 192 \procs. \begin{figure} @@ -597,4 +607,5 @@ \label{fig:locality:jax} \end{figure} + \begin{figure} \subfloat[][Throughput share]{ @@ -629,65 +640,34 @@ \end{figure} -Figures~\ref{fig:locality:jax} and \ref{fig:locality:nasus} show the results for the locality experiment. -In both cases, the graphs on the left column show the results for the @share@ variation and the graphs on the right column show the results for the @noshare@. -Looking at the left column on Intel first, Figures~\ref{fig:locality:jax:share:ops} and \ref{fig:locality:jax:share:ns}, which shows the results for the @share@ variation. -\CFA and Tokio slightly outperform libfibre, as expected based on their \ats placement approach. -\CFA and Tokio both unpark locally and do not suffer cache misses on the transferred array. -Libfibre on the other hand unparks remotely, and as such the unparked \at is likely to miss on the shared data. -Go trails behind in this experiment, presumably for the same reasons that were observable in the churn benchmark. -Otherwise the results are similar to the churn benchmark, with lower throughput due to the array processing. -As for most previous results, all runtime suffer a performance hit after 48 \proc, which is the socket boundary. - -Looking next at the right column on Intel, Figures~\ref{fig:locality:jax:noshare:ops} and \ref{fig:locality:jax:noshare:ns}, which shows the results for the @noshare@ variation. -The graph show the expected performance inversion where libfibre now outperforms \CFA and Tokio. -Indeed, in this case, unparking remotely means the unparked \at is less likely to suffer a cache miss on the array. -The leaves the \at data structure and the remote queue as the only source of likely cache misses. -Results show both are armotized fairly well in this case. +Looking at the right column on Intel, Figures~\ref{fig:locality:jax:noshare:ops} and \ref{fig:locality:jax:noshare:ns} show the results for the noshare variation. +The graphs show the expected performance inversion where libfibre now outperforms \CFA and Tokio. +Indeed, in this case, unparking remotely means the unparked \at is less likely to suffer a cache miss on the array, which leaves the \at data structure and the remote queue as the only source of likely cache misses. +Results show both are amortized fairly well in this case. \CFA and Tokio both unpark locally and as a result suffer a marginal performance degradation from the cache miss on the array. -Looking now at the results for the AMD architecture, Figure~\ref{fig:locality:nasus}, the results show a story that is overall similar to the results on the Intel. +Looking at the results for the AMD architecture, Figure~\ref{fig:locality:nasus}, shows results similar to the Intel. Again overall performance is higher and slightly more variation is visible. Looking at the left column first, Figures~\ref{fig:locality:nasus:share:ops} and \ref{fig:locality:nasus:share:ns}, \CFA and Tokio still outperform libfibre, this time more significantly. -This is expected from the AMD server, which has smaller and more narrow caches that magnify the costs of processing the array. -Go still sees the same poor performance as on Intel. +This advantage is expected from the AMD server with its smaller and more narrow caches that magnify the costs of processing the array. +Go still has the same poor performance as on Intel. Finally looking at the right column, Figures~\ref{fig:locality:nasus:noshare:ops} and \ref{fig:locality:nasus:noshare:ns}, like on Intel, the same performance inversion is present between libfibre and \CFA/Tokio. -Go still sees the same poor performance. - -Overall, this experiment mostly demonstrates the two options available when unparking a \at. +Go still has the same poor performance. + +Overall, this benchmark mostly demonstrates the two options available when unparking a \at. Depending on the workload, either of these options can be the appropriate one. -Since it is prohibitively difficult to detect which approach is appropriate, all runtime much choose one of the two and live with the consequences. - -Once again, this demonstrate that \CFA achieves equivalent performance to the other runtime, in this case matching the faster Tokio rather than Go which is trailing behind. +Since it is prohibitively difficult to dynamically detect which approach is appropriate, all runtimes much choose one of the two and live with the consequences. + +Once again, these experiments demonstrate that \CFA achieves equivalent performance to the other runtimes, in this case matching the faster Tokio rather than Go, which is trailing behind. \section{Transfer} The last benchmark is more of an experiment than a benchmark. It tests the behaviour of the schedulers for a misbehaved workload. -In this workload, one of the \at is selected at random to be the leader. +In this workload, one \at is selected at random to be the leader. The leader then spins in a tight loop until it has observed that all other \ats have acknowledged its leadership. The leader \at then picks a new \at to be the next leader and the cycle repeats. -The benchmark comes in two flavours for the non-leader \ats: +The benchmark comes in two variations for the non-leader \ats: once they acknowledged the leader, they either block on a semaphore or spin yielding. - -The experiment is designed to evaluate the short-term load-balancing of a scheduler. -Indeed, schedulers where the runnable \ats are partitioned on the \procs may need to balance the \ats for this experiment to terminate. -This problem occurs because the spinning \at is effectively preventing the \proc from running any other \at. -In the semaphore flavour, the number of runnable \ats eventually dwindles down to only the leader. -This scenario is a simpler case to handle for schedulers since \procs eventually run out of work. -In the yielding flavour, the number of runnable \ats stays constant. -This scenario is a harder case to handle because corrective measures must be taken even when work is available. -Note, runtime systems with preemption circumvent this problem by forcing the spinner to yield. - -In both flavours, the experiment effectively measures how long it takes for all \ats to run once after a given synchronization point. -In an ideal scenario where the scheduler is strictly FIFO, every thread would run once after the synchronization and therefore the delay between leaders would be given by: -$ \frac{CSL + SL}{NP - 1}$, where $CSL$ is the context switch latency, $SL$ is the cost for enqueueing and dequeuing a \at and $NP$ is the number of \procs. -However, if the scheduler allows \ats to run many times before other \ats are able to run once, this delay will increase. -The semaphore version is an approximation of the strictly FIFO scheduling, where none of the \ats \emph{attempt} to run more than once. -The benchmark effectively provides the fairness guarantee in this case. -In the yielding version however, the benchmark provides no such guarantee, which means the scheduler has full responsibility and any unfairness will be measurable. - -While this is a fairly artificial scenario, it requires only a few simple pieces. -The yielding version of this simply creates a scenario where a \at runs uninterrupted in a saturated system, and starvation has an easily measured impact. -However, \emph{any} \at that runs uninterrupted for a significant period of time in a saturated system could lead to this kind of starvation. +Figure~\ref{fig:transfer:code} shows pseudo code for this benchmark. \begin{figure} @@ -732,10 +712,39 @@ \end{figure} +The experiment is designed to evaluate the short-term load-balancing of a scheduler. +Indeed, schedulers where the runnable \ats are partitioned on the \procs may need to balance the \ats for this experiment to terminate. +This problem occurs because the spinning \at is effectively preventing the \proc from running any other \at. +In the semaphore variation, the number of runnable \ats eventually dwindles down to only the leader. +This scenario is a simpler case to handle for schedulers since \procs eventually run out of work. +In the yielding variation, the number of runnable \ats stays constant. +This scenario is a harder case to handle because corrective measures must be taken even when work is available. +Note, runtimes with preemption circumvent this problem by forcing the spinner to yield. + +In both variations, the experiment effectively measures how long it takes for all \ats to run once after a given synchronization point. +In an ideal scenario where the scheduler is strictly FIFO, every thread would run once after the synchronization and therefore the delay between leaders would be given by, $(CSL + SL) / (NP - 1)$, +where $CSL$ is the context-switch latency, $SL$ is the cost for enqueueing and dequeuing a \at, and $NP$ is the number of \procs. +However, if the scheduler allows \ats to run many times before other \ats are able to run once, this delay increases. +The semaphore version is an approximation of strictly FIFO scheduling, where none of the \ats \emph{attempt} to run more than once. +The benchmark effectively provides the fairness guarantee in this case. +In the yielding version however, the benchmark provides no such guarantee, which means the scheduler has full responsibility and any unfairness is measurable. + +While this is an artificial scenario, in real-life it requires only a few simple pieces. +The yielding version simply creates a scenario where a \at runs uninterrupted in a saturated system and the starvation has an easily measured impact. +Hence, \emph{any} \at that runs uninterrupted for a significant period of time in a saturated system could lead to this kind of starvation. + \subsection{Results} -\begin{figure} + +\begin{table} +\caption[Transfer Benchmark on Intel and AMD]{Transfer Benchmark on Intel and AMD\smallskip\newline Average measurement of how long it takes for all \ats to acknowledge the leader \at. +DNC stands for ``did not complete'', meaning that after 5 seconds of a new leader being decided, some \ats still had not acknowledged the new leader.} +\label{fig:transfer:res} +\setlength{\extrarowheight}{2pt} +\setlength{\tabcolsep}{5pt} \begin{centering} -\begin{tabular}{r | c c c c | c c c c } -Machine & \multicolumn{4}{c |}{Intel} & \multicolumn{4}{c}{AMD} \\ -Variation & \multicolumn{2}{c}{Park} & \multicolumn{2}{c |}{Yield} & \multicolumn{2}{c}{Park} & \multicolumn{2}{c}{Yield} \\ +\begin{tabular}{r | c | c | c | c | c | c | c | c} +Machine & \multicolumn{4}{c |}{Intel} & \multicolumn{4}{c}{AMD} \\ +\cline{2-9} +Variation & \multicolumn{2}{c|}{Park} & \multicolumn{2}{c |}{Yield} & \multicolumn{2}{c|}{Park} & \multicolumn{2}{c}{Yield} \\ +\cline{2-9} \procs & 2 & 192 & 2 & 192 & 2 & 256 & 2 & 256 \\ \hline @@ -746,44 +755,46 @@ \end{tabular} \end{centering} -\caption[Transfer Benchmark on Intel and AMD]{Transfer Benchmark on Intel and AMD\smallskip\newline Average measurement of how long it takes for all \ats to acknowledge the leader \at. -DNC stands for ``did not complete'', meaning that after 5 seconds of a new leader being decided, some \ats still had not acknowledged the new leader.} -\label{fig:transfer:res} -\end{figure} - -Figure~\ref{fig:transfer:res} shows the result for the transfer benchmark with 2 \procs and all \procs, where each experiment runs 100 \at per \proc. +\end{table} + +Table~\ref{fig:transfer:res} shows the result for the transfer benchmark with 2 \procs and all \procs on the computer, where each experiment runs 100 \at per \proc. Note that the results here are only meaningful as a coarse measurement of fairness, beyond which small cost differences in the runtime and concurrent primitives begin to matter. -As such, data points that are the on the same order of magnitude as each other should be basically considered equal. -The takeaway of this experiment is the presence of very large differences. +As such, data points within the same order of magnitude are basically considered equal. +That is, the takeaway of this experiment is the presence of very large differences. The semaphore variation is denoted ``Park'', where the number of \ats dwindles down as the new leader is acknowledged. The yielding variation is denoted ``Yield''. -The experiment was only run for the extremes of the number of \procs since the scaling is not the focus of this experiment. - -The first two columns show the results for the the semaphore variation on Intel. -While there are some differences in latencies, \CFA is consistenly the fastest and Tokio the slowest, all runtime achieve results that are fairly close. -Again, this experiment is meant to highlight major differences so latencies within $10\times$ of each other are considered close to each other. - -Looking at the next two columns, the results for the yield variation in Intel, the story is very different. -\CFA achieves better latencies, presumably due to the lack of synchronization on the semaphore. -Neither Libfibre or Tokio complete the experiment. -Both runtime use classical work-stealing scheduling and therefore since non of the work-queues are ever emptied no load balancing occurs. +The experiment is only run for many \procs, since scaling is not the focus of this experiment. + +The first two columns show the results for the semaphore variation on Intel. +While there are some differences in latencies, \CFA is consistently the fastest and Tokio the slowest, all runtimes achieve results that are fairly close. +Again, this experiment is meant to highlight major differences so latencies within $10\times$ of each other are considered equal. + +Looking at the next two columns, the results for the yield variation on Intel, the story is very different. +\CFA achieves better latencies, presumably due to no synchronization with the yield. +\todo{PAB: what about \CFA preemption? How does that come into play for your scheduler?} Go does complete the experiment, but with drastically higher latency: latency at 2 \procs is $350\times$ higher than \CFA and $70\times$ higher at 192 \procs. -This is because Go also has a classic work-stealing scheduler, but it adds preemption which interrupts the spinning leader after a period. +This difference is because Go has a classic work-stealing scheduler, but it adds coarse-grain preemption\footnote{ +Preemption is done at the function prolog when the goroutine's stack is increasing; +whereas \CFA uses fine-grain preemption between any two instructions.} +, which interrupts the spinning leader after a period. +Neither Libfibre or Tokio complete the experiment. +Both runtimes also use classical work-stealing scheduling without preemption, and therefore, none of the work queues are ever emptied so no load balancing occurs. Looking now at the results for the AMD architecture, the results show effectively the same story. The first two columns show all runtime obtaining results well within $10\times$ of each other. -The next two columns again show \CFA producing low latencies while Libfibre and Tokio do not complete the experiment. -Go still has notably higher latency but the difference is less drastic on 2 \procs, where it produces a $15\times$ difference as opposed to a $100\times$ difference on 256 \procs. - -This experiments clearly demonstrate that while the other runtimes achieve similar performance in previous benchmarks, here \CFA achieves significantly better fairness. -The semaphore variation serves as a control group, where all runtimes are expected to transfer leadership fairly quickly. -Since \ats block after acknowledging the leader, this experiment effectively measures how quickly \procs can steal \ats from the \proc running leader. -Figure~\ref{fig:transfer:res} shows that while Go and Tokio are slower, all runtime achieve decent latency. +The next two columns again show \CFA producing low latencies, while Go still has notably higher latency but the difference is less drastic on 2 \procs, where it produces a $15\times$ difference as opposed to a $100\times$ difference on 256 \procs. +Neither Libfibre or Tokio complete the experiment. + +This experiments clearly demonstrates \CFA achieves significantly better fairness. +The semaphore variation serves as a control, where all runtimes are expected to transfer leadership fairly quickly. +Since \ats block after acknowledging the leader, this experiment effectively measures how quickly \procs can steal \ats from the \proc running the leader. +Table~\ref{fig:transfer:res} shows that while Go and Tokio are slower using the semaphore, all runtimes achieve decent latency. + However, the yielding variation shows an entirely different picture. -Since libfibre and Tokio have a traditional work-stealing scheduler, \procs that have \ats on their local queues will never steal from other \procs. -The result is that the experiment simply does not complete for these runtime. -Without \procs stealing from the \proc running the leader, the experiment will simply never terminate. +Since libfibre and Tokio have a traditional work-stealing scheduler, \procs that have \ats on their local queues never steal from other \procs. +The result is that the experiment simply does not complete for these runtimes. +Without \procs stealing from the \proc running the leader, the experiment cannot terminate. Go manages to complete the experiment because it adds preemption on top of classic work-stealing. -However, since preemption is fairly costly it achieves significantly worst performance. +However, since preemption is fairly infrequent, it achieves significantly worst performance. In contrast, \CFA achieves equivalent performance in both variations, demonstrating very good fairness. -Interestingly \CFA achieves better delays in the yielding version than the semaphore version, however, that is likely due to fairness being equivalent but removing the cost of the semaphores and idle-sleep. +Interestingly \CFA achieves better delays in the yielding version than the semaphore version, however, that is likely due to fairness being equivalent but removing the cost of the semaphores and idle sleep.