Spring RTS Engine

an ati x1270 in my netbook, with an amd athlon 64 1.2 ghz cpu. Runs decently playable in 1v1 games, no rendering artifacts (~15 fps) all settings on lowest, resolution 1366x768.
Wouldnt cry if support for it would get dropped, its not for gaming.

CarRepairer wrote:Average FPS:
1-10

Max Units:
2

SpliFF wrote:

It's a netbook, I got it for web and other non-intensive tasks. I can spec small to medium games on spring but I don't expect much more than that.

More importantly, it is a VIA powered notebook.

In my opinion, dump support for the integrated stuff, you can get a geforce 7xxx series for 10 - 15 bucks and while not uber awesome, it will run.

==Troy== wrote:Because spring is nowhere near commercial games in gfx featureset? And the appeal of spring is different from the commercial games.

Hmm that's interesting, played Gundam, CA or Evo lately?

In some areas, our effects are better than some commercial games.

Forboding Angel wrote:In my opinion, dump support for the integrated stuff, you can get a geforce 7xxx series for 10 - 15 bucks and while not uber awesome, it will run.

The title of the thread says "laptops".
Shit Forb, get it together.

well technically you CAN upgrade a laptop GFX using external PCIe 1x but you won't get it for $10 and you still won't get great performance (1x bus). Still, like he said ^ I even repeated it in the OP.

jK wrote:I would like to drop the whole Intel/FFP support, most commercial games don't run on them why should spring?
-> minimum for the GPU: GLSL shaders (no fugly combiners anymore)

Btw even when you drop the FFP support, a new model renderer should be modularized enough to easily integrate new render methods. So ppl should be able to re-implement it themselves (with less features).

That wording sounds like dropping all Intel support. For reference, any relatively recent Intel chips has GLSL.

lurker wrote:

jK wrote:I would like to drop the whole Intel/FFP support, most commercial games don't run on them why should spring?
-> minimum for the GPU: GLSL shaders (no fugly combiners anymore)

Btw even when you drop the FFP support, a new model renderer should be modularized enough to easily integrate new render methods. So ppl should be able to re-implement it themselves (with less features).

That wording sounds like dropping all Intel support. For reference, any relatively recent Intel chips has GLSL.

huh?
Apple's drivers for Intel GPUs have GLSL, but ppl told me it's very crashy.
Also Mesa begins to add GLSL to their drivers, and because Intel's drivers build up on those, they claim they would do too. But this `support` is unfinished and has no hardware-acceleration.
And in the end some of Intel's newer GPUs use PowerVR technology which support GLSL, but afaik Intel is just too lazy to merge PowerVR's drivers into theirs.

Well, it works fine on my non-Apple system. Before I made that post I searched for a benchmark using GLSL, and furmark claims to. It runs with no CPU use, so I assume it's properly hardware accelerated. I'm not sure what point you're going for, other than Apple+Intel being bad, which I'll believe you about. Is support also broken on Linux? I haven't checked.

Nice to hear.

What GPU do you have btw?

GMA 4500MHD, roughly the same computer as manolo_

(HP Mini 5101)
* 1 GB 800 MHz DDR2 SDRAM
* Intel┬« AtomÔäó Processor N280 (1.66 GHz, 512 KB L2 cache, 533 MHz FSB)
* Mobile Intel® Graphics Media Accelerator 950

Spring settings on lowest, 3d trees, 24-bit buffer, res: 1366 x 758 => about 30 fps on default zoom in beginning vs bot, ie few units.

Problem was just that laptop has no middle mouse button :)

Oh, and mod was xta.

Laptop name: Azer Extensa 5630EZ
CPU: Intel ® Pentium ® Dual Core Processor T3400 (2.16GHz, 667Mhz FSB, 1MB L2 cache)
GFX accelerator: Up to 1244 MB Mobile Intel ® Graphics Media Accelerator 4500M
Resolution: 1280x800

--

Your spring settings:
Lowest. (i would use lowest even with epic super computer anyways)

Your spring view radius: (this is the setting that makes the most FPS for you.)
66.

Your average FPS:
At start: 50-60 on DSI with no units at all
250 flashes on screen: 24fps
500 flashes on screen: 14fps
750 flashes on screen: 10fps
1000 flashes on screen: 7fps

Max Units:
By looking at the FPS table, i would say 500 is enough for me to look at once.
But the CPU can handle more units, just the GPU cannot.
I tested, and each 1000 units lowers my FPS approximately by 5.
So if i had 50 fps at beginning with no units, i could spam 8000 units and still have 10 fps (if i dont look at any of the units at the screen at that time.
But if i look about 250 units at once and spam as many units as i can, then i would have only about 3000 units max to play with.

ETU:
Upgrade a shitty laptop? why? I buy shit because it doesnt cost much, so if it broke, i wouldnt cry. (and i barely ever use it so it doesnt need to be best ever you can get...)

CHEAP LAPTOPS are not for gaming!!!

This thread is pretty useless because of the fact above.

Forboding Angel wrote:For fps, next time you're in a decent sized BA game, just press B (or if you're running the new RC, alt+b).

IIRC this lowers your framerate a fair bit, doesn't it?

to see fps, just type /fps

It's even in the control panel, it's one of those things they put in before they decided to never, ever improve the control panel again.

IMHO we should really avoid developing for the lowest common denominator when hardware is progressing so fast, especially on the CPU front. Within a year, the first 32/28nm CPUs will be on the market and the first ones manufactured will be the simplest - this is the manufacturing paradigm that has been used for a while now. At the smaller node, less power will be used - but for example Atom CPUs already use only a few watts. Intel won't bother reducing the power usage any more than absolutely necessary, rather they'll respond to demand by making CPUs that are actually more on a par performance-wise with what you'd expect from a proper notebook. Simply put, the shrink to 28nm will bring a huge leap forward in terms of performance.

Similarly, the overwhelming demand for cheap GPUs that are capable of handling HD video decoding has brought NV into the market in the form of the Ion, to which Intel has already responded with a much more capable GMA chip. The shrink to 28nm will bring great benefits here too, but probably mostly for desktops since onboard video manufacturing tends to lag a generation or three behind so they can continue to make stuff on their old lines. NV's Fermi looks to be completely out of the running for the budget segment, so they are re-releasing their old cards with new names at extremely low prices. Even now ATI is producing GPUs that run Crysis nicely for $100, and this is before the 28nm shrink.

I'm all for broad hardware compatibility, but you have to draw the line somewhere. What I'm saying is, don't develop with GMA950 in mind. Even today at the "top of the line", GMA chipsets are intentionally designed to be obsolete so Intel has something to manufacture at its old plants. The situation is starting to improve quite quickly at the moment though. So please, don't develop with Geforce 4/FX in mind. Don't develop with Radeon 9000 or something in mind, either. Just say, "Your card must support pixel shader 3 and should have at least 128mb of VRAM". Or if you want to be cautious, say 256mb of VRAM. It is not reasonable for people to expect 10 year old GPUs to still run modern games, least of all now when the price for a top-notch GPU is so low. The only thing that makes me at all sympathetic is the AGP issue, but there are options like the Radeon 3850 which are very inexpensive.

I've come up with a plan regarding ongoing support for low-end graphics. Please feel free to comment in that thread. I don't need any more system specs here. Thanks to everyone who participated.

Just because the frontier of technology is marching forwards doesn't mean this high technology is automatically applied to 3 quarters of the market.

My friend jenny was working on support and she was shocked by the number of people who didn't have enough ram to start an empty java vm, who then claimed they had 32MB SDRAM and ti was all they could afford, living in the US may I add.

I believe that a piece of software that has to compromise itself to support the lowest denominater is not a project with a crazy project lead, but a project that has major design issues. A flexible engine should be modular enough to run on a wide variety of machines while pushing high end PCs.

If you feel like writing a 2D GUI/renderer the bar could be lowered more but obviously there are limits to how low you can go. I agree that allowing low-end machines to play is a worthy consideration but it isn't the most important thing either.

The BIGGEST hurdle to finding players is graphics quality. You can complain about how commercial games focus on this too much but the numbers prove it. A great looking game will outsell an entertaining game by at least 10-1. Gamers shop with their eye-balls (screen shots sell) and open-source doesn't change that.

The best example I can give is Rigs Of Rods. The concept is impressive - amazing even - but the graphics really blow, especially the male character you see when you're not in a vehicle. He basically glides and hops around like some kind of retarded ghost. It really put me off - it shouldn't - but it does. FreeCiv is another example. Its graphics are even reasonable but it looks a million years old next to Civ III.

Not saying we shouldn't try and support the low-end, just saying we shouldn't neglect the top-end to do it.