SupCom & Intel Core 2 Quad Gameplay Advantages

[Betrayer00GT]

My original plan was to get an e6300 for my old Intel motherboard to get it back up and running. After this review, I think I am going to wait until the prices drop a bit and get a q6600 for this machine, and move its e6600 into that Intel board. Yet more money, but it is all in good fun.

Also, as a plug for 64-bit Windows, I believe I got another of these crashes after the update to SupCom and Windows. SupCom itself was up to 2.3GB of mem as listed for its entry in the Processes tab. Being that I run Vista, I am assuming that everything "user", in total, caused me to bump against the new 3GB limit.

Also, after it crashed and I saw "supremecommander.exe" was using 2.3GB of RAM, I just restarted the game and reloaded a couple minute old save of that skirmish, which only used 1.4GB after it loaded and I could start playing. Then, after 15 minutes or so of real time, the RAM assigned to SupCom in the Processes tab was only up to 1.7GB. I finished the game really soon after that, and it didn't crash (or run the RAM usage up as high), again. I assume from this that SupCom is not a perfect manager of its own memory space. It seemed to be using a good bit extra for no real reason in the case of the skirmish that crashed.

 

[BobSutan]

Me and some guys at work talked about this basically all afternoon. The general consensus gravitated to the question of why is the game so bogged down? Yes, its advanced and all that, but not so graphically or else it would have experienced serious boosts from going ot the 8800SLI. That not being the case it demonstrates that the CPU is the bottleneck, but that doesn't make sense. If its not a poorly coded game that's causing the poor performance, what are we all missing?

 

[Dan_D]

Me and some guys at work talked about this basically all afternoon. The general consensus gravitated to the question of why is the game so bogged down? Yes, its advanced and all that, but not so graphically or else it would have experienced serious boosts from going ot the 8800SLI. That not being the case it demonstrates that the CPU is the bottleneck, but that doesn't make sense. If its not a poorly coded game that's causing the poor performance, what are we all missing?

RTS games have always been CPU intensive. That's why most games have a really low unit and building cap. Only more recent RTS's have abandoned the cap or at least made it insanely high. I would suspect that buy going to a quad core CPU that Supreme Commander could potentially improve alot. (I haven't read any of the articles on the game as I am not a big fan of most RTSs.)

 

[jebo_4jc]

yea supcom is pretty CPU limited. (hence the whole point of this thread!)

 

[StealthyFish]

it's only "needed" if you "need" to play this particular game (and maybe some of those others) at the high settings)

sounds like someone needs a higher end computer

 

[joggle]

In addition to running the utility that MadBoris provided to fix the 2gb memory limitation problem there are a couple of other things worth trying with these benchmarks. It would have been nice to set the resolution to the lowest setting for both Vista and XP for quad core and dual core (but with other CPU-influencing options at their highest) since this would ensure that the CPU is the only limiting factor (rather than some buggy Vista graphics driver). It also would have been nice to try popping in a Core Duo 2 of the same (or preferably higher) clock frequency. This would be a good way to verify that disabling two cores really is the same as running a single Core Duo 2. Since an Intel quad 4 is really just two Core Duo 2 processors sitting side by side this seems reasonable. And since processors with lower numbers of cores tend to clock faster than ones with more cores using a Core Duo 2 with a higher frequency seems fair, especially since it would still cost a lot less than a quad processor.

 

[PGHammer]

Recently STALKER has arrived and it claims the same support. I wonder how it would manage on these tests?

I must say these benchmarks are scary.. The game is far better with more cores... I might have to change what I get also...

I actually have a system-build designed for both x64 and multi-core (however, it's based on the Q6600, not QX6700), and it's outfitted with 8 GB of low-latency Mushkin. From the review at [H], SupCom will scale VERY well with the Q (even at the WS 1680x1050 typical resolution I'll be using). While quad-core is still expensive, the caveat is still *for now*. The review *also* points out that SupCom is more CPU-limited than you would think; the core scaling, as well as the poor single-core performance, are showing quite plainly that the only substitute for more cores is higher clocking (the poor single-core performance I would have to attribute to the clock-frequency deficit that even the stock E/QX/Q6700 will have compared to the midrange Pentium D 9xx, let alone the Prescott and Northwood-C S478 Netburst ancestors (which, while having a slower FSB, have a higher-clocked core than even the P-D).