Well, looks like Ryzen is perfect if you use gentoo, kek.

what about browser performance in linux

File: REKd.png (52KB, 720x289px) Image search: [Google]

52KB, 720x289px

IT'S HAPPENING

More Linux productivity benchmarks.

http://www.phoronix.com/scan.php?page=article&item=ryzen-1800x-linux&num=5

>the OS that doesn't receive bribes from intel has better performance for AMD products

huh really makes you think

>>59212606
Hope you're joking

>>59209279
>Use
>hobby OS
pick one

>>59212560
Productivity on Linux, nice one I needed a good laugh.

More importantly, does it work with Windows 7 at all?

>>59212606
You don't have to worry, clear as day Windows is simply garbage but thankfully enterprise and datacenter who buy chips don't give much of a damn about gaymer performance :)

File: told.png (134KB, 437x282px) Image search: [Google]

134KB, 437x282px

>>59209279

File: reading.jpg (123KB, 1280x720px) Image search: [Google]

123KB, 1280x720px

>>59209279

gcc options: -lavdevice -lavfilter -lavformat -lavcodec -lswresample -lswscale -lavutil -lXv -lX11 -lXext -lxcb -lxcb-xfixes -lxcb-render -lxcb-shape -lasound -lSDL -lm -llzma -lbz2 -pthread -std=c99 -fomit-frame-pointer -fPIC -O3 -fno-math-errno -fno-signed-zeros -fno-tree-vectorize

Why does he never set -march=native?

>>59210911
Wat? "Browser performance"?

You did realize some low power ARM processors can handle browsers well already, right?

This is a much more power hungry but also much more performant desktop CPU, you don't have to worry about the CPU side of browsers.

this settles it, AMD has won on /g/

>>59209279
This is the only shit that matters.

>>59216570
Clearly you haven't seen the 640x480 gaming benchmarks which are a much better indicator of realworld CPU performance

>>59216583 [citation needed]

>>59212540
What do you know, the OS that doesn't have an intel compiler does better.

File: 1356046333522.jpg (14KB, 288x344px) Image search: [Google]

14KB, 288x344px

>>59209279
twf still not twice as fast as my 7 year old 980x

>>59212560
>>59213207
>polish intellectuals

>>59216655
Moore's law is dead for CPUs.

>>59216707
it's a 32nm CPU ffs

You know that a dual Xeon would smoke that all day long, right? And typing make -c 57 would just make me smile.

>>59217779

You know that quad Xeon would smoke that all day long, right?

>hurr thrown more hardware at the problem that is fixed with more hardware

No shit

>>59213150
http://www.dailytech.com/FTC+Intel+Reach+Settlement+Intel+Banned+From+Anticompetitive+Practices/article19265.htm

>>59217796
Solid argument, but $500 for a ryzen, and go to eBay to get a dual ES e52670v4, now you have 40 threads, for $100 less, and will run circles around ryzen all day long, while staying cooler, and more stable in the process.

>>59218271
Used hardware is higher value than new hardware.
Call the press.

Nobody should ever upgrade to new hardware anymore since old hardware is much higher in perf/$

>>59218292
Get two brand new Xeon e52620's, 24 threads, still smoke ryzen. Are you done moving goalposts yet?

>>59218311
The motherboard alone costs more than the entire Ryzen setup.

>>59209279
it's a fucking compute monster

why does it suck so bad at gaming?

>>59218311
> Processor Base Frequency 2.00 GHz
>Max Turbo Frequency 2.50 GHz
>$400 per pop, without dual socket motherboards

Lol, not only does the Ryzen demolish it in single threaded performance like a P4, it demolishes it in everything else while using less power.

Used hardware is shit, enjoy your PCIe 2.0 and USB 2.0 support.

>>59218333
$500 cpu + $135 mobo = ryzen setup
$150 * 2 CPU + $300 mono = dual Xeon

It doesn't matter how far you move the posts, cause at the end of the day, ryzen is still going to be no good for this application.

>>59218384
>$500 CPU
>when the $400 performs 2% slower and the $300 one performs 5% slower

Lol, nice cherrypicking, it still stands that Ryzen demolishes it in single thread performance and unless your workload scales 100% with cores(doesn't), your two housefires with overpriced motherboards on a 2010 chipset without native PCIe 3.0 suport gets raped in its kidneys.

User hardware is worthless unless your wallet is worthless.

>>59218412
Why are we talking about single core performance now? We're compiling a kernel. Get lost, shill.

>>59218412
You don't need pcie to compile Linux, shill. And the $300 being %5 slower, while the dual Xeon is still 10x faster? Have you ever compiled anything? More cores absolute speeds things up, tremendously.

>>59218478
All you do is compile a kernel? Even the most parallel tasks has serial parts to its workload.
And it's basic math, if your core is twice slower your double core count can't compete without 100% scaling, which is impossible.

>>59218509
Quit pretending to be retarded idiot, compiling on 1 4ghz is way slower than 2 2ghz cores, cause the 2 core can make 2 files at once, while the single can only make one at a time. And I do more than compile, but the thread is about compiling a kernel, you idiot, so don't change the subject just to get an advantage.

>>59218384

A whole brand new system with used refurbished parts that not even sure if it will work?

>>59209279

OP, i'm also interested in installing it. I'm about to get a job at costco. (first job) I assume that i'm going to have co-workers, that i'm occasionally going to have to talk to. I don't want to be the boring guy who has nothing going on in his life. So, when someone says, "Tell me something interesting about yourself", I'd like to be able to respond with something like "Well, I'm one of the gentoomen", or something like that. But I can't exactly do that if I haven't actually installed gentoo.

>>59218551
Non refurbished parts, brand new. Look up prices if you don't believe me.

HURRRRRRRRR Anonymous 04/28/07(Sat)21:57 No.321086
Man you guys are retarded. Gentoo isn't 'hard' or 'useless' and portage isn't 'laughable' if you actually know what you're doing. If you haven't used Linux for very long, use Debian or something.

I'm not trying to sound pretentious here, but giving someone who doesn't know Linux a gentoo install and expecting them not to break it is like giving your 16 year old kid a ferrari and hoping he doesn't speed and run off the road.

-----

That sounds gay. I'm going to pronounce it with a hard G like in Great, en as in enforce, too as in too.

File: 51997.png (39KB, 575x500px) Image search: [Google]

39KB, 575x500px

>>59218531
>1-2 cores
Nice cherrypicking, the higher the core count is the harder it is to parallelize it.


>muh mor weak corz is better
No it's not, you're literally arguing for Bulldozer.

Get fucked, shill.

>>59218384
You can get a shitty used OEM motherboard for $150, a decent new one costs $400. With cheap server boards/oems you might run into problems and wasting more money at the end of the day.

I'll wait some time until compiler optimizations/memory/clock/ecc and other issues get solved and look back. The current benchmarks seem not taking the best out of Ryzen and some results don't make sense at all.

File: 67413k.png (32KB, 650x525px) Image search: [Google]

32KB, 650x525px

>>59218600

>>59218600
>>59218688
Notice that they're all around the same clock speed, the main difference is the number of cores. You don't seem to understand how this works, but try to follow.

>>59218614
>compiler optimizations
Do AMD fags blame EVERYTHING on this? Fuck off.

File: 52688.png (41KB, 575x500px) Image search: [Google]

41KB, 575x500px

>>59218688
Literally the only workload where it does well.

Bulldozer was shit in servers because its IPC was so low, it's high integer performance (eg compiling linux 1 second faster doing nice in hashes) wasn't enough to make it relevant against Ivybridge

>>59218746
I don't understand your argument, but it sounds like you're for Xeon, but not as a desktop processor, which I'm fine with.

>>59218766
I'm against low IPC and ST performance in general, and 2.0GHz Sandybridge is lacking in both, even a 7700k will run around it in circles at just 4 cores not to mention a 6900k or a Ryzen that run around the 7700k in circles.

tldr

10 cores at 2.0Ghz is worse than 4 cores at 4.5GHz or 8 cores at 3.5GHz

Massively so.

>>59218803
Whatever you say, bro.

>>59218746
>SAP
>used for benchmarks

hahaha what the fuck. it's probably the ugliest shittiest code base in the world.

>>59218828
Unfortunately it's a very real life workload, Adobe crap is uglier than sin and it's used for benching "productivity" frequently.

>>59218828
>As of 2016, SAP is the world's third largest software and programming company.
Regardless of its ugliness it's still very important, not that I don't agree with you, but that's reality.