• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

Question How fast is your browser?

Page 7 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Magic Carpet

Diamond Member
Oct 2, 2011
3,336
120
106
@Det0x

5950x IS stupidly fast, almost on par with Apple's M1 Safari. Makes me want to buy it, lol. Too bad, you can't bench it with the Safari browser, unless you do this.

@general
Make sure, edge://gpu, chrome://gpu is accelerated and extensions disabled for the best score. In my case, uBlock alone can bring the score down by around 10%. Also, older computers with iGPU like Sandy/Ivy running Windows 10 tend to perform a bit worse in this test than with Windows 7/8 but that's likely due to a different driver model, not fully supported by those GPUs. Windows 10 performs best with WDDM 2.0 and up.

1618584211376.png1618584521837.pngnnkeoh69fuz51.png
 
Last edited:
  • Like
Reactions: lightmanek

Det0x

Senior member
Sep 11, 2014
513
645
136
127.6
Windows 10 19042.928
Firefox 87


Ok. I've got the same CPU. What clockspeed are you running this? Impressive.
Everything is tuned.. PBO CO at -30 allcore together +100 mhz offset + custom watercooling and with very tight memory timings.
With that 1677points ST in Cinebench r23 i would guestimate i was running around 5100 mhz effective clock
Only easy trick i found was to stop the browser benchmark bouncing between CCD's.. Forcing Chrome to run at thread 0-15 made me go from 200points to 211points.

Also Google Chrome Version 90.0.4430.72 in Incognito mode seems to be the fastet..
 

Makaveli

Diamond Member
Feb 8, 2002
4,311
502
126
Everything is tuned.. PBO CO at -30 allcore together +100 mhz offset + custom watercooling and with very tight memory timings.
With that 1677points ST in Cinebench r23 i would guestimate i was running around 5100 mhz effective clock
Only easy trick i found was to stop the browser benchmark bouncing between CCD's.. Forcing Chrome to run at thread 0-15 made me go from 200points to 211points.

Also Google Chrome Version 90.0.4430.72 in Incognito mode seems to be the fastet..
At PBO C0 -20 on my two fastest cores and -10 + 200 mhz on the rest has me at 202.6 in standard mode not using incognito there is no difference for me in this mode as I have no ad blockers or add ons installed, all my ad blocking is done at the router level in my setup.

The Edge Version 91.0.862.0 (Official build) canary (64-bit)

In performance mode scores 193.1 so abit faster than the 189 score I was getting on Version 89.0.774.77 (Official build) (64-bit)
 
Last edited:

Gideon

Golden Member
Nov 27, 2007
1,353
2,588
136
Cool, any chance you can run the same test on Safari?
Here you go:

Screenshot 2021-05-06 at 11.35.29.png

Didn't bother to run in incognito as I don't use that browser. That's with full dev environment, IDE's, chat apps, roughly (10 microservices) humming in the backround in Docker (true they had no load), etc ... while the fan is not starting up.

But yeah this is apples to oranges. Would really like to see how a hackintosh with a Ryzen 5xxx series would perform. At least in theory @Det0x 's rig should be able to hit ~250
 
Last edited:

john3850

Golden Member
Oct 19, 2002
1,418
17
81
Does does your ip speed have much to do any score.
I also found that WebKit team at Apple made it.
 
Last edited:

Magic Carpet

Diamond Member
Oct 2, 2011
3,336
120
106
But yeah this is apples to oranges. Would really like to see how a hackintosh with a Ryzen 5xxx series would perform. At least in theory @Det0x 's rig should be able to hit ~250
I heard, that Safari also takes advantage of JS HW acceleration that M1 has. If you compare battery life on battery power, Safari should be way more efficient than Chromium. Obviously, the more energy-efficient libraries/code help tremendously. Still, it would be nice to compare it with 5950x in Hackintosh mode.


Does does your ip speed have much to do any score.
I get similar results whether I'm on a 1gb Ethernet or slower Wi-Fi. It shouldn't affect your score. CPU is responsible for the vast chunk of the score, but not alone obviously, GPU and web browser also play a part. Try running it with a basic GPU driver for example.

I also found that WebKit team at Apple made it.
Yeah, that's no news.
 

john3850

Golden Member
Oct 19, 2002
1,418
17
81
I got 1 number higher with basic GPU driver
The GPU isn't actually much faster, if at all, then the CPU, but it carries out the instruction set at the same time, thus saving, in affect, time.
Thanks for your input Magic Carpet
 
Last edited:

JoeRambo

Golden Member
Jun 13, 2013
1,119
979
136
Any TGL-H scores from the web, with browser that is easy to compare with some results in the web? (not of the kind "we restored our golden image with IE4 we use for web testing since 1999"). Really interested in how those large Intel L2/L3 caches perform in this.
 
  • Like
Reactions: Magic Carpet

ASK THE COMMUNITY