11 Replies. 1 pages. Viewing page 1.
Newer [  1  ] Older
11.
 
Re: Saturday Interviews
Sep 8, 2014, 05:47
11.
Re: Saturday Interviews Sep 8, 2014, 05:47
Sep 8, 2014, 05:47
 
Hmm, Tom, that was quite technical. You wouldn't happen to be this Tom, would you?
If Russia stops fighting, the war ends. If Ukraine stops fighting, Ukraine ends. Slava Ukraini!
Avatar 22024
10.
 
Re: Saturday Interviews
Sep 7, 2014, 14:17
Tom
10.
Re: Saturday Interviews Sep 7, 2014, 14:17
Sep 7, 2014, 14:17
Tom
 
Um yeah, vsync does cap framerates. Let's start with the basics. Vsync stands for vertical synchronization, which means synchronizing the updating of the framebuffer of a graphics card with the vertical blanking interval of a CRT. A CRT contains an electron gun which is sweeping over the display to refresh the phosphors that maintain the image on the screen. During the vertical blanking interval, the framebuffer can safely be updated without causing tearing. Tearing happens when you see part of an old frame along with part of a new frame because the framebuffer got updated while the monitor was in the middle of refreshing the visible contents of the screen.

Technology has moved on since the days of CRTs but vsync still means syncing framebuffer updates with the refresh rate of the display device. If you're performing such synchronization, you aren't updating the framebuffer more frequently than that. Hence the refresh rate serves as a limiting factor (cap) on the rate at which frames are rendered and presented to the user.

Of course, there are plenty of ways for a programmer to be an idiot and screw this up, but really - what sense is there in asking the GPU to render more frames than the display will pass along to the user? It's just wasteful and stupid.

You are correct in that a frame limiter is different from vsync. A frame limiter simply caps the framerate with no attempt at synchronizing to the rate of something else such as the display's refresh process. This is why you get tearing if you use a frame limiter set to your monitor's refresh rate and don't use vsync. There is no synchronization.
9.
 
Re: Saturday Interviews
Sep 7, 2014, 13:27
9.
Re: Saturday Interviews Sep 7, 2014, 13:27
Sep 7, 2014, 13:27
 
Tom wrote on Sep 7, 2014, 11:22:
It's sad that they manage to do this even when the framerate is capped by vsync.

Vsync does not cap framerates, it is something entirely different from a frame limiter (which works BEFORE the frame-buffer while vsync works by limiting frames already passed to the frame-buffer (by preventing them from reaching it)) Meaning you need both, and fyi, frame limiter at 60fps would still lead to tearing ;p
Avatar 54727
8.
 
Re: Saturday Interviews
Sep 7, 2014, 11:22
Tom
8.
Re: Saturday Interviews Sep 7, 2014, 11:22
Sep 7, 2014, 11:22
Tom
 
The overheating thing shouldn't really be surprising to anyone who has gamed on a non-entry-level smartphone before. These things can get seriously toasty. And you can kiss your battery goodbye.

The good news is that as mobile processors continue to get more powerful, they should get more useful at lower levels of utilization. We can see this effect with desktop GPUs. If I lower the temperature target on my 780s to 70C, I can get nice cool and quiet performance even when inept game developers have irresponsibly programmed their games to waste all available resources no matter what. It's sad that they manage to do this even when the framerate is capped by vsync.
7.
 
Re: Saturday Interviews
Sep 7, 2014, 09:50
Sho
 
7.
Re: Saturday Interviews Sep 7, 2014, 09:50
Sep 7, 2014, 09:50
 Sho
 
jdreyer wrote on Sep 7, 2014, 03:20:
So your phone has four cores, but you can't USE four cores. At least not in anything more than a very short burst.

It depends on your definition of "using". In practice a CPU core will often go idle briefly even while a process is technically using it (e.g. because it's waiting on I/O), and power management then works to let it descend into various sleep states that e.g. turn parts of the chip off via clock gating to have it draw less power - and thus also leak less heat. A situation where all four cores are at C0 (not sleeping) for any length of time is going to be rare in practice. And even at meager utilization multi-core can have performance benefits, especially because mobile CPU cores are traditionally not really designed for ILP.

That said, I'm sure there are offenders in the smartphone industry who push down on the thermal envelope too aggressively.

This comment was edited on Sep 7, 2014, 09:59.
6.
 
Re: Saturday Interviews
Sep 7, 2014, 03:20
6.
Re: Saturday Interviews Sep 7, 2014, 03:20
Sep 7, 2014, 03:20
 
I also thought this was really interesting:

The other thing that was a real surprise to me was how important the power management was, not just for battery life, but for performance. Because if you go ahead and fire up all four cores at the top clock rate and draw a bunch of stuff, you overheat within minutes. If you really load everything onto the system that can draw power, you will probably thermally throttle down in under a minute. So you literally just can’t use all the capabilities of the system at the same time.

So your phone has four cores, but you can't USE four cores. At least not in anything more than a very short burst.
If Russia stops fighting, the war ends. If Ukraine stops fighting, Ukraine ends. Slava Ukraini!
Avatar 22024
5.
 
Re: Saturday Interviews
Sep 7, 2014, 03:01
5.
Re: Saturday Interviews Sep 7, 2014, 03:01
Sep 7, 2014, 03:01
 
There is plenty of anecdotal evidence that overuse of a cell phone (especially the older ones with much large SAR values) can be dangerous. An inordinate number of floor traders constantly sending in orders via cell phone ended up with brain tumors on their favored side. This was in the 90s. And also in the 90s, one of my students' fathers was a telephone lineman who used a first gen wireless radio telephone to check the line he just set. He started to get migraines, and they found four separate brain tumors on his right side where he'd used the telephone.

I'll grant you that cell phones broadcast less radiation than those. But why take the chance? I usually use a headset when I call someone on my cell. I usually leave my cell in my pack, not in my pocket near my balls. Simple precautions aren't inconvenient.

The one thing I'll say about VR is this: radio radiation dissipates exponentially from the broadcast source. I'd be much less worried about having the cell phone in a visor 3 inches from my face than I would pressing it up against my ear for a phone call. Also, if you were worried about the radiation, you could put the phone into airplane mode while you had it in the visor.
If Russia stops fighting, the war ends. If Ukraine stops fighting, Ukraine ends. Slava Ukraini!
Avatar 22024
4.
 
Re: Saturday Interviews
Sep 7, 2014, 01:51
4.
Re: Saturday Interviews Sep 7, 2014, 01:51
Sep 7, 2014, 01:51
 
Sho wrote on Sep 6, 2014, 18:55:
DangerDog wrote on Sep 6, 2014, 17:45:
Cell phone radiation is a real thing

[citation needed]

http://www.cnet.com/pictures/highest-cell-phone-radiation/

The SAR value of cell phones is usually hidden from consumers, at least in the US anyway.
Avatar 6174
3.
 
Re: Saturday Interviews
Sep 6, 2014, 19:41
3.
Re: Saturday Interviews Sep 6, 2014, 19:41
Sep 6, 2014, 19:41
 
DangerDog wrote on Sep 6, 2014, 17:45:
Cell phone radiation is a real thing, I don't think I would want to wear a cell phone on my head for hours on end playing games/watching movies like that.
Of course it's real. You wouldn't be able to see anything on the screen if there were no radiation :-)
2.
 
Re: Saturday Interviews
Sep 6, 2014, 18:55
Sho
 
2.
Re: Saturday Interviews Sep 6, 2014, 18:55
Sep 6, 2014, 18:55
 Sho
 
DangerDog wrote on Sep 6, 2014, 17:45:
Cell phone radiation is a real thing

[citation needed]
1.
 
Re: Saturday Interviews
Sep 6, 2014, 17:45
1.
Re: Saturday Interviews Sep 6, 2014, 17:45
Sep 6, 2014, 17:45
 
Cell phone radiation is a real thing, I don't think I would want to wear a cell phone on my head for hours on end playing games/watching movies like that.
Avatar 6174
11 Replies. 1 pages. Viewing page 1.
Newer [  1  ] Older