PDA

View Full Version : Php 5.1.0


orban
Thu 24th Nov '05, 7:22pm
Anybody compiled that yet? :) Tempting, they claim a lot of performance improvements.

kerplunk
Fri 25th Nov '05, 12:55am
We're using it with APC 3.0.8. There is a HUGE speed difference between PHP 5.0 and PHP 5.1. Zend Engine 2 is very nice. I would strongly recommend upgrading to PHP 5.1.

orban
Fri 25th Nov '05, 4:55am
Nice to hear!

Does it work with eAccelerator 0.9.4-rc1 (http://sourceforge.net/mailarchive/forum.php?thread_id=9027313&forum_id=43418)?

eva2000
Fri 25th Nov '05, 5:05am
kerplunk, how's speed between PHP 5.1.0 vs PHP 4.4.1 ?

DaFire
Fri 25th Nov '05, 7:43am
Nice to hear!

Does it work with eAccelerator 0.9.4-rc1 (http://sourceforge.net/mailarchive/forum.php?thread_id=9027313&forum_id=43418)?
according to the eaccelerator forum - nope.

I think I give APC a try... what are the differences to eaccelerator ?

orban
Fri 25th Nov '05, 9:56am
I'm interested in that too. I used eAcc in the past and it really works/worked great.

DaFire
Fri 25th Nov '05, 10:02am
i try to make some benchmarks on my development box .. I want to test:

php 4.4.1 (without accelerator)
php 4.4.1 eaccelerator
php 4.4.1 apc
php 5.1 (without accelerator)
php 5.1 apc

It's a small box ( i think a 600 mhz amd duron or something like that running freebsd ) but it should be ok for comparison.

I plan to use apache ab on the same box for benchmarking.

orban
Fri 25th Nov '05, 10:53am
Apache 2 or 1 and PHP compiled as apxs?

DaFire
Fri 25th Nov '05, 11:05am
ok .. I made a few tests now:

Apache 2.0.55
./configure --enable-so

php 4.4.1 & php 5.1.0
./configure --with-apxs2=/usr/local/apache2/bin/apxs --with-mysql=/usr --with-gd --with-zlib-dir=/usr/lib

I called the benchmark tool with

/usr/local/apache2/bin/ab -n 100 localhost/forum/index.php
(100 requests)

I ran all tests twice after restarting the server to make sure the caches are already filled

Results:

php 4.4.1 (no accelerator)

Concurrency Level: 1
Time taken for tests: 66.805268 seconds
Complete requests: 100
Failed requests: 10
(Connect: 0, Length: 10, Exceptions: 0)
Write errors: 0
Total transferred: 4040667 bytes
HTML transferred: 4001467 bytes
Requests per second: 1.50 [#/sec] (mean)
Time per request: 668.053 [ms] (mean)
Time per request: 668.053 [ms] (mean, across all concurrent requests)
Transfer rate: 59.05 [Kbytes/sec] received

Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 0 0.0 0 0
Processing: 650 667 8.4 667 695
Waiting: 609 625 7.8 625 650
Total: 650 667 8.4 667 695

Percentage of the requests served within a certain time (ms)
50% 667
66% 669
75% 670
80% 670
90% 672
95% 689
98% 691
99% 695
100% 695 (longest request)

php 5.1 (no accelerator)

Concurrency Level: 1
Time taken for tests: 69.248684 seconds
Complete requests: 100
Failed requests: 64
(Connect: 0, Length: 64, Exceptions: 0)
Write errors: 0
Total transferred: 4025118 bytes
HTML transferred: 3985918 bytes
Requests per second: 1.44 [#/sec] (mean)
Time per request: 692.487 [ms] (mean)
Time per request: 692.487 [ms] (mean, across all concurrent requests)
Transfer rate: 56.75 [Kbytes/sec] received

Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 0 0.0 0 0
Processing: 671 691 11.8 695 725
Waiting: 625 650 11.8 654 684
Total: 671 691 11.8 695 725

Percentage of the requests served within a certain time (ms)
50% 695
66% 698
75% 700
80% 702
90% 704
95% 711
98% 718
99% 725
100% 725 (longest request)

php 4.4.1 (APC)

Concurrency Level: 1
Time taken for tests: 43.317657 seconds
Complete requests: 100
Failed requests: 10
(Connect: 0, Length: 10, Exceptions: 0)
Write errors: 0
Total transferred: 4040667 bytes
HTML transferred: 4001467 bytes
Requests per second: 2.31 [#/sec] (mean)
Time per request: 433.177 [ms] (mean)
Time per request: 433.177 [ms] (mean, across all concurrent requests)
Transfer rate: 91.07 [Kbytes/sec] received

Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 0 0.1 0 1
Processing: 415 432 7.4 434 460
Waiting: 397 414 7.5 416 442
Total: 415 432 7.4 434 460

Percentage of the requests served within a certain time (ms)
50% 434
66% 436
75% 437
80% 438
90% 439
95% 439
98% 443
99% 460
100% 460 (longest request)

php 5.1 (APC)

Concurrency Level: 1
Time taken for tests: 42.574035 seconds
Complete requests: 100
Failed requests: 4
(Connect: 0, Length: 4, Exceptions: 0)
Write errors: 0
Total transferred: 4031688 bytes
HTML transferred: 3992488 bytes
Requests per second: 2.35 [#/sec] (mean)
Time per request: 425.740 [ms] (mean)
Time per request: 425.740 [ms] (mean, across all concurrent requests)
Transfer rate: 92.47 [Kbytes/sec] received

Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 0 0.1 0 1
Processing: 402 424 8.2 427 454
Waiting: 379 401 8.0 403 430
Total: 402 424 8.2 427 454

Percentage of the requests served within a certain time (ms)
50% 427
66% 428
75% 429
80% 430
90% 431
95% 432
98% 433
99% 454
100% 454 (longest request)


My conclusion after the first test is that plain php5.1 is a bit slower than php 4.4.1 but using apc its slightly faster.

orban
Fri 25th Nov '05, 12:45pm
eAcc coming soon? :D

eAcc with PHP 5.0.5 and 4.xx that is.

And thanks a lot!

Scott MacVicar
Fri 25th Nov '05, 1:22pm
Use APC, its written by one of the core PHP developers (who works for Yahoo where its also used)

Tantawi
Fri 25th Nov '05, 4:00pm
I upgraded too and it works great with APC 3.0.8.

orban
Fri 25th Nov '05, 10:29pm
I really have to check out APC then :)

BamaStangGuy
Sat 26th Nov '05, 1:23am
I really have to check out APC then :)
APC user here as well. :)

eva2000
Sat 26th Nov '05, 5:31am
ok .. I made a few tests now:

My conclusion after the first test is that plain php5.1 is a bit slower than php 4.4.1 but using apc its slightly faster.
excellent comparison there :)

DaFire
Sat 26th Nov '05, 8:58am
I have seen benchmarks where the difference between php4/php5 and php5.1 are huge ... php5.1 a few times faster then the older ones but I guess in a real life situation ( the page i benchmarked is forum frontpage with vbadvanced cmps boxes on the left side ) the things that count are database server and those still run with the same speed :)

version2
Mon 28th Nov '05, 3:06pm
Performance-wise...does it make sense to go ahead and move on to php5?

kerplunk
Mon 28th Nov '05, 5:13pm
Performance-wise...does it make sense to go ahead and move on to php5?
It appears so. We're running PHP 5.1.1 with APC and it's awesome. :)