Closed Bug 410197 Opened 17 years ago Closed 16 years ago

Download speed of files slower than firefox 2.

Categories

(Toolkit :: Downloads API, defect)

x86
Windows Vista
defect
Not set
minor

Tracking

()

VERIFIED WORKSFORME

People

(Reporter: itshectic, Unassigned)

References

()

Details

(Keywords: perf)

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9b2) Gecko/2007121120 Firefox/3.0b2
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9b2) Gecko/2007121120 Firefox/3.0b2

Speedtest.net results:
Firefox 2: 35000
Firefox 3: 30000

Microsoft.com download:
Firefox 2: 4100kb/s 
Firefox 3: 3.6MB/s

Reproducible: Always

Steps to Reproduce:
1. Use windows Vista
2. Have Firefox 2 and Firefox 3 both installed
3. Switch between the 2 using 1 browser at a time to test download speed (did not have both running at the same time)
Actual Results:  
Consistant overall download speed is much slower on new version of firefox.
New version of firefox never reaches the capable download speed of version 2.

Expected Results:  
Downloaded at the same speed.

Internet explorer is also slow at downloading.

Maximum speed of download in windows XP is 30Mbit when using firefox 2.

Currently I am using Vista Ultimate 32bit which has increased my download speed to 35Mbit as seen in firefox 2
I did also a download test with Windows XP on www.speedtest.nl, repeated it quite some times, but Firefox 3 was even a bit faster in this test. 
I never suspected this bug to affect XP.
Vista has TCP recieve window auto-tuning which I think may be related to this bug.
I did also a test with Vista, which is indeed a bit slower.
Each round I did the test 10 times, 3 rounds totally.

branch: 1105+927+893+954+963+860+933+948+964+945=9492/10=949kB
trunk: 909+909+955+927+931+864+938+820+954+922=9129/10=912kB

branch: 901+904+954+968+972+979+926+963+957+966=9490/10=949kB
trunk: 888+926+919+651+643+909+843+912+793+887=8371/10=837kB

branch: 742+989+950+929+695+770+565+818+669+689=7816/10=781kB
trunk: 908+909+875+870+929+946+848+936+945+936=9102/10=910kB
As you can see in bug 407928 it should be happen with a high bandwidth. I'm not able to test it at the moment.
Version: unspecified → Trunk
Please retest with a currently nightly?

Bug 410131 switched to buffered writes, which probably ought to help here.
I run a test with Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9b3pre) Gecko/2008010813 Minefield/3.0b3pre ID:2008010813 for the given URL:

Download: 84561 kb/s
Upload:   64101 kb/s

Using Firefox 2.0.0.11 gives me much slower download rates:

Download: 45569 kb/s
Upload:   60215 kb/s

It would be interesting to test with an older trunk build before the patch on bug 410131 was checked-in. I'll do this later.
You may want to confirm those numbers by testing a few times, alternating FF versions. If there's a transparent proxy, load balancer, etc somewhere on the connection you could see the first browser tested as slow, and the next one fast.
I repeated the tests several times for each version. Nothing changes for Fx2. It only has a maximum download rate of about 45000 kb/s. Also the ping to the server is higher as with Fx3.

VERSION   DOWNLOAD_SPEED  UPLOAD_SPEED  LATENCY  SERVER_NAME  DISTANCE
Firefox3  "78311 kb/s"    "68389 kb/s"  "8 ms"   "Frankfurt"  "~ 100 mi"
Firefox3  "84561 kb/s"    "34101 kb/s"  "6 ms"   "Frankfurt"  "~ 100 mi"
Firefox2  "44770 kb/s"    "57039 kb/s"  "15 ms"  "Frankfurt"  "~ 100 mi"
Firefox2  "45569 kb/s"    "60215 kb/s"  "10 ms"  "Frankfurt"  "~ 100 mi"
Firefox3  "87261 kb/s"    "64008 kb/s"  "5 ms"   "Frankfurt"  "~ 100 mi"
Firefox2  "46298 kb/s"    "40019 kb/s"  "15 ms"  "Frankfurt"  "~ 100 mi"
Firefox2  "41203 kb/s"    "61121 kb/s"  "12 ms"  "Frankfurt"  "~ 100 mi"
Firefox3  "91324 kb/s"    "61808 kb/s"  "7 ms"   "Frankfurt"  "~ 100 mi"
Firefox3  "87425 kb/s"    "65759 kb/s"  "6 ms"   "Frankfurt"  "~ 100 mi"

Looks interesting. Do we really have such a download speed boost with Firefox 3 now?
I did a real test with downloading Ubuntu from a server next to me. As I can see again Firefox 3 is clearly faster as Firefox 2.

Firefox 2: ~6MB/s
Firefox 3: ~10MB/s

These values show that the measured data from my last comment are ok. Even saving the file to disk doesn't slow down Firefox 3.

From my side this bug is WFM. Reporter, can you still see your issue?
(In reply to comment #10)
> I did a real test with downloading Ubuntu from a server next to me. As I can
> see again Firefox 3 is clearly faster as Firefox 2.
> 
> Firefox 2: ~6MB/s
> Firefox 3: ~10MB/s
> 
> These values show that the measured data from my last comment are ok. Even
> saving the file to disk doesn't slow down Firefox 3.
> 
> From my side this bug is WFM. Reporter, can you still see your issue?
> 
thu jan 10 2008
I have today tested twice both the minefield and internet explorer browsers on my PC. Results as follows: IE 22.4MB/s and Minefield 21.92MB/s For both max speed attained downloading a 100MB file saved to disk. <speedtest.bbned.nl>
Seems the issue has vanished, thanks.
Jacob

Sweet - I guess that buffering patch helped out a lot. :)
Status: UNCONFIRMED → RESOLVED
Closed: 16 years ago
Resolution: --- → WORKSFORME
(In reply to comment #12)
> Sweet - I guess that buffering patch helped out a lot. :)

I'm not really sure. I tested a build from 1st january - before the patch was checked-in - and I didn't realize a difference in the download speed. No idea if there were some other improvement fixes.
(In reply to comment #13)
> (In reply to comment #12)
> > Sweet - I guess that buffering patch helped out a lot. :)
> 
> I'm not really sure. I tested a build from 1st january - before the patch was
> checked-in - and I didn't realize a difference in the download speed. No idea
> if there were some other improvement fixes.
>

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9b3pre) Gecko/2008011105 Minefield/3.0b3pre
Confused...
I have timed the minefield download manager with my mobile telefone stopwatch.
a--The average download time for a 100MB file saved to desktop at 2.74MB/s (internet connection speed 23MB at time of test) is around and about 31 seconds.
b--minefield download manager  download time is (starts at 50 seconds) unreal and the app does n`t show the elapse time for the specific download.
c--besides the above the app is doing well with regards to speed and time it should reflects resp. true/real. I hope it is clear.
jacob 
Resolution: WORKSFORME → FIXED
Resolution: FIXED → WORKSFORME
Status: RESOLVED → VERIFIED
Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9b4pre) Gecko/2008021704 Minefield/3.0b4pre
OS WXP-h  internet conn. 24MB
Downloaded this morning ubuntu 8.04 desktop live-cd 697MB in just 0:05:02.09
The download time shown by the minefield download Manager is more realistic now.
In my opinion the minefield browser and download Manager is doing great.
jacob
Product: Firefox → Toolkit
You need to log in before you can comment on or make changes to this bug.