AMD's Display Core difficulties

Posted Dec 14, 2016 16:40 UTC (Wed) by pizza (subscriber, #46)
In reply to: AMD's Display Core difficulties by Cyberax
Parent article: AMD's Display Core difficulties

Because they are used for different things?

(Oh, for example, A server sitting in a rack without any heads acting as a NAS on a trusted network has different needs than graphically-intensive EDA tools on a portable workstation that might connect to a public hotspot)

AMD's Display Core difficulties

Posted Dec 14, 2016 16:41 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link] (7 responses)

So why would one kind of a server need slower network?

AMD's Display Core difficulties

Posted Dec 14, 2016 16:56 UTC (Wed) by zdzichu (subscriber, #17118) [Link] (6 responses)

Slower network performance comes as a drawback of more strict firewall configuration and lower power usage. Probably. There is no analysis in this benchmark, neither “perf” nor any other tracing to pinpoint bottlenecks.

AMD's Display Core difficulties

Posted Dec 14, 2016 17:10 UTC (Wed) by Cyberax (✭ supporter ✭, #52523) [Link] (5 responses)

Nope. Linuxes tested by default have zero firewall rules.

That's why Phoronix test are useful - they make people to at least look at this stuff. Tests with more details would be nice, but in their current form they are quite thought-provoking.

AMD's Display Core difficulties

Posted Dec 14, 2016 17:42 UTC (Wed) by pizza (subscriber, #46) [Link]

> Nope. Linuxes tested by default have zero firewall rules.

Incorrect. I can't speak for the others, but Fedora 25 (and I believe CentOS7) default to having a firewall enabled, even on minimal installations.

> That's why Phoronix test are useful - they make people to at least look at this stuff. Tests with more details would be nice, but in their current form they are quite thought-provoking.

No.. they really aren't. But with a bit more work on their part, they could be.

AMD's Display Core difficulties

Posted Dec 14, 2016 18:11 UTC (Wed) by johannbg (guest, #65743) [Link] (3 responses)

You can get the same performance out of the same components running on the same kernel on the same hardware regardless what that distribution would be called so benchmarking tests from moronix is just comparing default generic settings, settings which are determined by the community surrounding each distribution to cater to that communities needs and maintainability thus are pretty much useless.

Truly benchmarking, tweaking, tuning and, identifying the bottleneck and report where they resides is a very time consuming process.

Those moronix benchmarks just remind me of those idiots that brag about their uptime on hw with no true sustained load over period of time.

Bragger: Duh my *nix box has been running for 400 days!
Me: Has it now, show me the sustained load on the machine?
Bragger: load average: 0.01, 0.01, 0.01
Me: so basically you have just had that machine turned on for 400 days consuming power and paying for it so you can brag for it's uptime to other people. Bragger you must be a special kind of idiot.

Moronix is just another echo chamber of pour if that site can be called journalism on the internet much like mashdot.

This seems like a good stopping point

Posted Dec 14, 2016 18:27 UTC (Wed) by corbet (editor, #1) [Link] (2 responses)

If you have a technical issue with somebody's benchmarking techniques (or whatever), those can be legitimate items for discussion — though it's kind of off-topic here. But please can we refrain from childish name calling? That doesn't help anybody and doesn't make the conversation any more fun.

Anyway, my feeling is that this thread has pretty much run its course.

This seems like a good stopping point

Posted Dec 14, 2016 19:01 UTC (Wed) by johannbg (guest, #65743) [Link] (1 responses)

Dont stop on my account my feeling is that this thread is still full of rainbow, joy and ponies and "childish name calling" is in line with the quality of the technical journalism that takes place on both of those sites so arguably it fits perfectly so no I cant see why I or others should stop themselves from calling it so. ( And no I did not come up with those names myself, this is the nick name these sites have on these parts )

If each comment is required to help somebody than you can look at it this way.
People actually can get dumber ( through misinformation ) from reading articles from those sites so just view that comment as an public service announcement for this sites readers own health and I would be doing them a disservice not pointing that out.

Now whether that triggers readers emotion response such as joy I cannot say since each emotional response varies between each individual reader behind his or her screen so meh.

This seems like a good stopping point

Posted Dec 14, 2016 20:45 UTC (Wed) by rahulsundaram (subscriber, #21946) [Link]

> Dont stop on my account my feeling is that this thread is still full of rainbow, joy and ponies and "childish name calling" is in line with the quality of the technical journalism that takes place on both of those sites so arguably it fits perfectly so no I cant see why I or others should stop themselves from calling it so.

Kinda hard to criticize someone for doing something if you are "in line" with it.