{"id":4246,"date":"2016-08-20T16:04:29","date_gmt":"2016-08-20T15:04:29","guid":{"rendered":"https:\/\/stevepedwards.today\/DebianAdmin\/?p=4246"},"modified":"2016-08-20T16:04:29","modified_gmt":"2016-08-20T15:04:29","slug":"how-slow-can-you-go-old-and-new-tech-moores-law-comparison-2-program-performance","status":"publish","type":"post","link":"https:\/\/stevepedwards.today\/DebianAdmin\/how-slow-can-you-go-old-and-new-tech-moores-law-comparison-2-program-performance\/","title":{"rendered":"How Slow Can You Go? Sysbench Comparisons – CPUs"},"content":{"rendered":"
<\/div>\n

<\/path><\/svg><\/i> \"Loading\"<\/p>\n

<\/div>\n

To continue on with this old V newer tech performance theme, as it is important to have references for at least 3 areas of system processes as indicators of performance: CPU; memory and disk IO.<\/p>\n

I have discovered Sysbench since reading Brendan Gregg's book, so will use that across my 4 different PCs, as it also standardises testing for each system by running the same task.<\/p>\n

sudo apt-get install sysbench<\/span><\/p>\n

NAME<\/span>
\n sysbench - A modular, cross-platform and multi-threaded benchmark tool.<\/span><\/p>\n

Sysbench can run some preset tests for various processes and their performance such as disk IO, mem IO etc.<\/p>\n

I am running a CPU, prime numbers based test on all 4 PCs for this Post - it should be self explanatory.<\/p>\n

All systems have the same\u00a0Mint kernel versions:\u00a03.16.0-38-generic, but different hardware and memory amounts as stated in the first Post for this theme:<\/p>\n

For each system, the results are summarised below with execution time being the result of interest, to give more or less linear ratios for performance ability, summarised at the end and basic conclusions drawn as for the first Post.<\/p>\n

HP AMDA8 1GHz Quad Core laptop, 8GB of RAM<\/strong><\/p>\n

stevee@AMDA8 ~ $ sudo sysbench --test=cpu --cpu-max-prime=100000 run<\/span>
\n[sudo] password for stevee: <\/span>
\nsysbench 0.4.12: multi-threaded system evaluation benchmark<\/span><\/p>\n

Running the test with following options:<\/span>
\nNumber of threads: 1<\/span><\/p>\n

Doing CPU performance benchmark<\/span><\/p>\n

Threads started!<\/span>
\nDone.<\/span><\/p>\n

Maximum prime number checked in CPU test: 100000<\/span>
\nTest execution summary:<\/span>
\n total time: 218.2423s<\/span>
\n total number of events: 10000<\/span>
\n total time taken by event execution: 218.2331<\/strong><\/span>
\n per-request statistics:<\/span>
\n min: 20.49ms<\/span>
\n avg: 21.82ms<\/span>
\n max: 29.82ms<\/span>
\n approx. 95 percentile: 25.00ms<\/span><\/p>\n

Threads fairness:<\/span>
\n events (avg\/stddev): 10000.0000\/0.00<\/span>
\n execution time (avg\/stddev): 218.2331\/0.00<\/strong><\/span><\/p>\n

HP Pavilion 4450e AMD Dual Core 64, 2300MHz, 66MHz mem bus, 4GB ram, SATA<\/strong><\/p>\n

stevee@hpmint ~ $ sudo sysbench --test=cpu --cpu-max-prime=100000 run<\/span>
\nsysbench 0.4.12: multi-threaded system evaluation benchmark<\/span><\/p>\n

Running the test with following options:<\/span>
\nNumber of threads: 1<\/span><\/p>\n

Doing CPU performance benchmark<\/span><\/p>\n

Threads started!<\/span>
\nDone.<\/span><\/p>\n

Maximum prime number checked in CPU test: 100000<\/span>
\nTest execution summary:<\/span>
\n total time: 867.1377s<\/span>
\n total number of events: 10000<\/span>
\n total time taken by event execution: 867.1276<\/strong><\/span>
\n per-request statistics:<\/span>
\n min: 86.61ms<\/span>
\n avg: 86.71ms<\/span>
\n max: 151.73ms<\/span>
\n approx. 95 percentile: 86.72ms<\/span><\/p>\n

Threads fairness:<\/span>
\n events (avg\/stddev): 10000.0000\/0.00<\/span>
\n execution time (avg\/stddev): 867.1276\/0.00<\/strong><\/span><\/p>\n

Dell Inspiron 531 Dual Core AMD 64, 2800MHz, 66MHz mem bus, 6GB ram, SATA<\/strong><\/p>\n

stevee@DELLMINT ~ $ sudo sysbench --test=cpu --cpu-max-prime=100000 run<\/span>
\nsysbench 0.4.12: multi-threaded system evaluation benchmark<\/span><\/p>\n

Running the test with following options:<\/span>
\nNumber of threads: 1<\/span><\/p>\n

Doing CPU performance benchmark<\/span><\/p>\n

Threads started!<\/span>
\nDone.<\/span><\/p>\n

Maximum prime number checked in CPU test: 100000<\/span>
\nTest execution summary:<\/span>
\n total time: 711.6993s<\/span>
\n total number of events: 10000<\/span>
\n total time taken by event execution: 711.6894<\/strong><\/span>
\n per-request statistics:<\/span>
\n min: 71.11ms<\/span>
\n avg: 71.17ms<\/span>
\n max: 101.45ms<\/span>
\n approx. 95 percentile: 71.20ms<\/span><\/p>\n

Threads fairness:<\/span>
\n events (avg\/stddev): 10000.0000\/0.00<\/span>
\n execution time (avg\/stddev): 711.6894\/0.00<\/strong><\/span><\/p>\n

HP Pavilion Intel Celeron M 1 Core 32, 1729MHz, 66MHz mem bus, 1GB ram, SATA<\/strong><\/p>\n

stevee@localhost ~ $ sudo sysbench --test=cpu --cpu-max-prime=100000 run<\/span>
\nsysbench 0.4.12: multi-threaded system evaluation benchmark<\/span><\/p>\n

Running the test with following options:<\/span>
\nNumber of threads: 1<\/span><\/p>\n

Doing CPU performance benchmark<\/span><\/p>\n

Threads started!<\/span>
\nDone.<\/span><\/p>\n

Maximum prime number checked in CPU test: 100000<\/span>
\nTest execution summary:<\/span>
\n total time: 644.3397s<\/span>
\n total number of events: 10000<\/span>
\n total time taken by event execution: 644.3194<\/strong><\/span>
\n per-request statistics:<\/span>
\n min: 53.37ms<\/span>
\n avg: 64.43ms<\/span>
\n max: 173.93ms<\/span>
\n approx. 95 percentile: 97.64ms<\/span><\/p>\n

Threads fairness:<\/span>
\n events (avg\/stddev): 10000.0000\/0.00<\/span>
\n execution time (avg\/stddev): 644.3194\/0.00<\/strong><\/span><\/p>\n

Well, this is surprising!<\/p>\n\n\n\n\n\n\n<\/colgroup>\n\n\n\n\n\n
<\/td>\nAMDA8 localhost core<\/strong><\/td>\nDell 64 bit dual core<\/strong><\/td>\nHP Pav 64 bit dual\u00a0<\/strong><\/td>\nHP Pav 32 bit mono<\/strong><\/td>\n<\/tr>\n
\u00a0Memory DRAM<\/td>\n\u00a08GB<\/td>\n\u00a06GB<\/td>\n\u00a04GB<\/td>\n\u00a01GB<\/td>\n<\/tr>\n
Time elapsed (secs)<\/strong><\/td>\n218.2331<\/strong><\/span><\/td>\n711.6894<\/strong><\/span><\/td>\n867.1276<\/strong><\/span><\/td>\n644.3194<\/strong><\/span><\/td>\n<\/tr>\n
Ratio to localhost core<\/td>\n\u00a0218\/<\/strong>218 = 1<\/strong><\/span><\/td>\n711\/218 =\u00a03.26<\/strong><\/span><\/td>\n867\/218 = 3.977\u00a0<\/strong><\/span><\/td>\n\u00a0644\/218 =\u00a02.95<\/strong><\/span><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n\n\n\n\n\n<\/colgroup>\n\n\n\n\n\n
\u00a0Added 4\/10\/16<\/strong><\/span><\/td>\nHP \u00a0Intel Pent Dual<\/strong><\/span><\/td>\n<\/td>\n<\/td>\n<\/td>\n<\/tr>\n
\u00a0Memory DRAM<\/strong><\/td>\n\u00a02GB<\/span><\/td>\n<\/td>\n<\/td>\n<\/td>\n<\/tr>\n
Time elapsed (secs)<\/strong><\/td>\n216.4174<\/span><\/strong><\/td>\n<\/td>\n<\/td>\n<\/td>\n<\/tr>\n
Ratio to localhost core<\/strong><\/td>\n\u00a0218\/216 = 0.991<\/span><\/strong><\/td>\n<\/td>\n<\/td>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

The Intel Dual is quicker than the AMD localhost for this task!<\/p>\n

The old 32 bit Celeron is NOT the slowest processor for this task by a long margin: in 2nd place - 67 secs before the Dell64 and 223 seconds before the HP64. It was 426 seconds behind the laptop, compared to the nearest rival Dell at 493 secs behind.<\/p>\n

I know there are other factors involved like level 1, 2 and 3 cache, DRAM make and clock speed, Front Side Bus and motherboard design, what other Apps are running etc. but it goes to show that 64 bit does not necessarily mean faster\/better depending...<\/p>\n

Going by these results for the worst case HP64 comparison, the localhost core was 867\/218 = 3.97 times faster,\u00a0and as only a single thread was run, that should be single core bound - as I understand that - so it suggests that a SINGLE CORE of the localhost core IS up to 4 times faster than a single core of any of the other CPUs.<\/p>\n

https:\/\/en.wikipedia.org\/wiki\/Thread_(computing)<\/a><\/span><\/p>\n

\"In computer science<\/a>, a thread of execution is the smallest sequence of programmed instructions that can be managed independently<\/strong> by a\u00a0scheduler<\/a>, which is typically a part of the operating system<\/a>.[1]<\/a><\/sup> The implementation of threads and processes<\/a> differs between operating systems, but in most cases a thread is a component of a process. Multiple threads can exist within one process, executing concurrently<\/a> and sharing resources such as memory<\/a>, while different processes do not share these resources. In particular, the threads of a process share its executable code and the values of its variables at any given time.<\/em><\/p>\n

Systems with a single processor generally implement multithreading by time slicing<\/a>: the central processing unit<\/a> (CPU) switches between different software threads. This context switching<\/a> generally happens very often and rapidly enough that users perceive the threads or tasks as running in parallel. On a multiprocessor<\/a> or multi-core<\/a> system, multiple threads can execute in parallel<\/a>, with every processor or core executing a separate thread simultaneously;<\/strong> on a processor or core with hardware threads<\/a>, separate software threads can also be executed concurrently by separate hardware threads.\"<\/em><\/p>\n

This is not in line with Moore's Law as a generalisation this time - as this localhostrupling\u00a0of processing power is over about a ten year period here, not 2 years.<\/p>\n

Considering this test and the prior Post on this, Moore's Law seems to apply loosely to overall systems, not necessarily individual components of type - taking ALL the system's performance capability into account i.e. where the slowest of all the components is the bottleneck that needs most consideration for the approximate claimed doubling of performance per halving in size every 2 years, as well the efficiency of the software if running software based tests for benchmarking.<\/p>\n

Moore's Law is a rough generalisation at best (and it's not a law as a Physics definition anyway) mainly describing manufacturing transistor density over time, as a general indicator of processing power.<\/p>\n