Request: Performance Measurements

[Fillip H.]

Hey all,

We're interested in hearing how DBSEO is performing on your site, and in order to get some more details on it we'd like it if you could take a few minutes minutes to measure this for us


If you aren't moving from vBSEO
Step 1: With the mod activated, go here: WebPagetest - Website Performance and Optimization Test and run it on your forum.php page
Step 2: Disable the mod via the Plugin Manager, and undo the .htaccess edits
Step 3: Go back to WebPageTest and re-run the tests from step 1
Step 4: Paste your result URLs in to this thread


If you are moving from vBSEO
Step 1: With vBSEO activated and DBSEO deactivated both on the forum and in the .htaccess, go here: WebPagetest - Website Performance and Optimization Test and run it on your forum.php page
Step 2: Disable vBSEO via the Plugin Manager, and undo the .htaccess edits
Step 3: Go back to WebPageTest and re-run the tests from step 1
Step 4: Disable vBSEO via the Plugin Manager, re-enable DBSEO and change the .htaccess back to the DBSEO rules
Step 5: Go back to WebPageTest and re-run the tests from step 1
Step 4: Paste your result URLs in to this thread


Please ensure you properly label the results so we know which is which. Thank you!

 

[janslu]

test server tests:

full loading time is dependent on various ads and variable...

vbseo

Document Complete	Fully Loaded
Load Time	First Byte	Start Render	Speed Index	DOM Elements	Time	Requests	Bytes In	Time	Requests	Bytes In
First View	9.599s	1.379s	3.811s	5113	5115	9.599s	132	892 KB	10.833s	142	917 KB
Repeat View	5.332s	1.937s	2.651s	4977	5102	5.332s	37	274 KB	7.435s	38	359 KB


none

		Document Complete	Fully Loaded
Load Time	First Byte	Start Render	Speed Index	DOM Elements	Time	Requests	Bytes In	Time	Requests	Bytes In
First View	6.381s	1.121s	2.519s	3581	6015	6.381s	130	832 KB	7.225s	139	858 KB
Repeat View	4.810s	0.915s	2.072s	2736	6031	4.810s	35	181 KB	4.951s	35	183 KB



dbseo

Document Complete	Fully Loaded
Load Time	First Byte	Start Render	Speed Index	DOM Elements	Time	Requests	Bytes In	Time	Requests	Bytes In
First View	6.734s	1.514s	2.911s	5751	6037	6.734s	135	937 KB	7.897s	147	964 KB
Repeat View	6.274s	1.678s	2.388s	3432	6014	6.274s	41	187 KB	7.473s	41	271 KB

 

[Fillip H.]

Thanks

Any way you could normalise the tests by removing the variable ads from your test server?

 

[janslu]

I have changed back to a stock vbulletin 4 style and have disabled all ads. But the problem is that the results are still very incostistent:

DBSEO:



	Load Time	First Byte	Start Render	Speed Index	DOM Elements	Time	Requests	Bytes In	Time	Requests	Bytes In
1.
First View	5.488s	1.826s	2.459s	2500	5670	5.488s	37	232 KB	6.042s	38	326 KB
Repeat View	5.008s	3.150s	3.396s	3650	5670	5.008s	3	72 KB	5.149s	3	74 KB
2.
First View	3.344s	1.840s	2.665s	2702	5671	3.344s	37	205 KB	3.878s	39	244 KB
Repeat View	2.528s	1.607s	2.092s	2102	5671	2.528s	6	76 KB	2.586s	6	78 KB
3.
First View	6.864s	3.417s	4.900s	5004	5670	6.864s	36	205 KB	8.984s	38	326 KB
Repeat View	2.927s	2.051s	2.413s	2400	5670	2.927s	7	88 KB	3.051s	7	90 KB



None:
1.
First View	3.285s	0.983s	1.689s	1700	5668	3.285s	35	203 KB	3.538s	36	225 KB
Repeat View	1.842s	0.883s	1.486s	1500	5668	1.842s	2	68 KB	1.983s	2	70 KB
2.
First View	3.245s	0.935s	1.643s	1600	5668	3.245s	35	205 KB	3.477s	36	226 KB
Repeat View	2.948s	1.270s	2.156s	2200	5668	2.948s	4	72 KB	3.167s	4	74 KB
3.
First View	3.122s	0.875s	1.553s	1600	5668	3.122s	34	204 KB	3.375s	35	226 KB
Repeat View	1.759s	0.818s	1.425s	1400	5668	1.759s	1	68 KB	2.300s	1	69 KB


VBSEO:

1. 
First View	4.385s	1.845s	2.632s	3062	4752	4.385s	37	221 KB	5.597s	38	325 KB
Repeat View	3.605s	2.436s	2.727s	3066	4752	3.605s	3	70 KB	3.731s	3	72 KB
2. 
First View	7.992s	5.583s	6.347s	6762	4752	7.992s	37	293 KB	8.149s	38	325 KB
Repeat View	2.864s	1.807s	2.048s	2416	4751	2.864s	2	70 KB	3.088s	2	71 KB
3.
First View	3.808s	1.421s	2.148s	2562	4752	3.808s	39	212 KB	5.794s	40	325 KB
Repeat View	3.187s	1.409s	1.782s	2247	4751	3.187s	5	74 KB	3.421s	6	76 KB

I'm afraid webpagetest is not really dependable in my case. The variability might be caused by a variable load on the main site or the database server. All tests were done in a span of 20 minutes...

 

[Fillip H.]

Could be, yeah. Thanks for going through the effort

 

[janslu]

One more questions - debug mode in vbulletin reports render time (along with a number of queries etc). Does it contain dbseo processing time? Maybe this might be a way to compare the performance independently of external factors?

 

[Fillip H.]

Unfortunately not, as part of both vBSEO and DBSEO's processing occurs before vBulletin is initialised.

 

[Mental]

One thing i'll throw in as a recommendation, dbseo seems to run smoother on xcache as opposed to apc (what i used for vbseo). I had to lower my xcache size and increased my var size, still yet to find a suitable configuration, atm i'm using the following;

;[xcache]
xcache.shm_scheme = "mmap"
xcache.size = 384M
xcache.count = 8
xcache.slots = 8K
xcache.ttl = 0
xcache.gc_interval = 300

xcache.var_size = 164M
xcache.var_count = 8
xcache.var_slots = 8K
xcache.var_ttl = 0
xcache.var_maxttl = 0
xcache.var_gc_interval = 300

xcache.test = Off
xcache.readonly_protection = Off
xcache.mmap_path = "/dev/zero"
xcache.coredump_directory = ""
xcache.cacher = On
xcache.stat = On
xcache.optimizer = Off

[xcache.coverager]

 

[CharlieDelta]

Disabled

http://www.webpagetest.org/result/131207_42_R4V/

Enabled

http://www.webpagetest.org/result/131207_YB_R8R/

 

[neounix]

FYI, we are seeing just about the same performance using DBSEO as using VBSEO for URL rewriting.

When we removed VBSEO, our performance improved by about 50% (for example, 250 ms GoogleBot download time became around 125ms).

Now that we have DBSEO running in place of VBSEO, the "GoogleBot download time" has returned to the same approximate performance level as when we had VBSEO in place.

I'm not saying this to be negative. We like DBSEO. I'm simply reporting our observation/experience that DBSEO URL rewriting and VBSEO URL rewriting both performance nearly the same on our server, as far a Googlebot downloading a page performance goes.

I'm a fan of DBSEO, BTW; and also was a fan of VBSEO before they stopped supporting the product. Now, we are really unhappy with the VBSEO owners; and we are very appreciative that DragonByte is supporting a product for vBulletin forums.

Note: We use memcached (obtw)

 

[RedWS600]

I did have VBSEO but I uninstalled it a couple of months ago, so I'm doing the test with DBSEO and no SEO.

With DBSEO

												Document Complete	Fully Loaded
			Load Time		First Byte	Start Render	DOM Elements		Time	Requests	Bytes In	Time	Requests	Bytes In
1.
First View		2.517s		0.365s		0.000s	1747				2.517s	37	810 KB	3.238s	47	847 KB
Repeat View	1.987s		0.309s		0.000s	1747				1.987s	8	410 KB	2.455s	10	437 KB
2.
First View		2.480s		0.324s		0.000s	1747				2.480s	38	793 KB	3.240s	48	845 KB
Repeat View	2.549s		0.299s		0.000s	1747				2.549s	13	458 KB	2.997s	15	495 KB
3.
First View		3.134s		0.818s		0.000s	1755				3.134s	39	795 KB	5.729s	49	920 KB
Repeat View	2.210s		0.318s		0.000s	1753				2.210s	6	405 KB	2.681s	7	410 KB

DBSEO OFF

												Document Complete	Fully Loaded
			Load Time		First Byte	Start Render	DOM Elements		Time	Requests	Bytes In	Time	Requests	Bytes In
1.
First View		2.605s		0.299s	0.000s		1753				2.605s	38	790 KB	3.216s	47	815 KB
Repeat View	2.368s		0.274s	0.000s		1753				2.368s	9	448 KB	2.849s	11	474 KB
2.
First View		2.776s		0.343s	0.000s		1753				2.776s	40	814 KB	5.844s	49	918 KB
Repeat View	2.240s		0.273s	0.000s		1753				2.240s	10	462 KB	2.731s	11	467 KB
3.
First View		2.460s		0.347s	0.000s		1753				2.460s	37	779 KB	3.221s	47	832 KB
Repeat View	1.897s		0.317s	0.000s		1753				1.897s	7	405 KB	2.358s	8	410 KB

 

[rtiindia]

Server performance good but mysql and Memory load increased

There are tremendous changes in the server performance, though apparently the site doesn't seem to be faster. Here are the plesk report. The DBSEO was installed on thursday night.

Out of all the reports, zombies process has increased tremendously, and I think that is the reason once I restart my server the speed becomes ok. Similarly, CPU usage by MySQL services has increased phenomenally.

 

[Elijah]

This one was taken a couple of weeks ago
WebPagetest Test Result - London : www.dreadsuk.com - 06/16/14 11:47:54

and this one was today after the install. Seems to be worse
WebPagetest Test Result - Dulles : www.dreadsuk.com - 07/08/14 21:13:16

 

[Fillip H.]

Going from no SEO to DBSEO will be a detriment to performance, that much is given, considering the amount of rewriting it needs to do

 

[Elijah]

Looking better now you've fixed it for me WebPagetest Test Result - London : www.dreadsuk.com - 07/09/14 21:19:49

 

[gscimbom]

WebPagetest Test Result - Dulles : www.gsc...om/forum/forum.php - 09/06/14 10:30:26 Do you have any suggestions?

 

[Fillip H.]

Get rid of all scripts and images that load from realist.gen.tr

 

[gscimbom]

Get rid of all scripts and images that load from realist.gen.tr

Thansk Fillip H.
New test results: WebPagetest Test Result - Dulles : www.gsc...om/forum/forum.php - 09/06/14 11:53:03

 

[Fillip H.]

reklamport.com, ca.cubecdn.com and every domain below #95 on the list found on that URL is now an outlier.

I haven't seen your forum, but if this list is to be believed you have a gigantic amount of ads on your page, which is going to have a severe negative impact on performance.

 

[tavenger5]

I didn't do a webpagetest but it is apparent in the Google crawl stats when I installed dbseo

Note that my attachments went from being hosted on a CDN to the server - it's a photography site so there are a lot of pictures. The attachment extensions were also not right and then corrected, so google image search is doing a lot of recrawling. Traffic is actually a bit down because of this, but seems to be picking back up since correcting the error.