StarCraft II Revisited: How Much Gaming PC Do You Need?

Page 5 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
Sweet. So they haven't started making any games that completely overpower my mid-ranged $1,300 3 year old machine. I can live without seeing a few things in high resolution.
 
Out of curiosity, what did you get for 1,300.00 three years ago?

I haven't picked this game up but I guess I will if it ever goes on sale. I never played the first so I don't have the fever, but I'd like to know what's so great about it.
 
Poor article, just like the previous TH article on SC2 benchmarks. Quite simply they don't remove the cpu bottleneck (i7 920 oc to a measly 3ghz???) and all you get are graphs where half the gfx cards are showing the same performance...

Yeah you can increase the res so you finally see some spread, but what about those gamers that want to see how their pc would/should perform with certain configurations? Like slightly lower res of 1680x1050, and a GTX460 paired with a decently overclocked (3.6ghz or higher) i5 750. The way this article is set up, such setups look like they perform almost the same as if you threw in a 5870 instead.
 
[citation][nom]Shirosaki[/nom]It's not CPU intensive at all.Had a 2.4 ghz Intel core 2 duo, 2 gigs of ddr2 ram, and a 460 1 gb gtx, WinXP 32 bit and I was getting 40-60 FPS in Ultra. I upgraded to I5 750, 4 gigs of ram, Win7 64 bit, Same GPU, and I get the same FPS as I did with the old set up. 40-60. There may be a few FPS difference I didn't write the exact numbers down. But it was still in the 40-60 range in the places I did pay attention to. Thought my old setup would be bottlenecking my GPU. Guess not.[/citation]
Would like to see some solid numbers backed with tested evidence than just huge ranges like 40-60fps.
 
I'm sure you guys would like to see some solid numbers, but I wasn't benchmarking the game for an article. I just didn't notice a performance increase with the new setup. And the range was more about the FPS being in the 40s at certain points and 60s at other points in the game and noticing the same thing with the new setup.

I also have my I5 OCed to 3.6 ghz. Forgot to mention that in the previous post. I'm sure that its more noticeable for people who fill a large map up with units attacking each other. I'm sure that is more CPU intensive, but how many gamers are going to play the game that way? None

And I didn't notice a FPS increase in my personal experience when upgrading. I did notice a huge difference in performance, but that was due to going from 2 to 4 gigs of ram. That makes a huge difference in this game.

I just think people should keep things in perspective, and know that the game isn't going to be as CPU intensive as suggested in this article. It's being benchmarked here, and it's pushed a lot further than average SC2 games are going to push it. One of the reasons I upgraded was because of SC2 and that was a huge mistake. I could have just bought 2 more gigs of ram and been fine for 6 more months until new processors and GPUs were out. I'd rather not see anyone else make the same mistake as me. I did not see a difference between a 2.4 ghz core2duo and a 3.6 ghz I5-750 in average games of SC2.
 
I probably shouldn't have said it's not CPU intensive at all, but should have been more clear about there being not as much difference in FPS performance between a 2.4ghz core2duo and a 3.6ghz I5-750 in average games. I meant more for average games. This Benchmark obviously proves that you can put the CPU under more load with a map full of units attacking each other. I wanted to edit my previous message last night but there doesn't appear to be an edit button for these comments.
 
THis is a real GAME benchmark article. Of course thanks to gigabyte, but it makes for an almost complete read except for the fact that heat or cpu noise/fan/power issues could have also been discussed on a page. But maybe next time.

 


That was the point though.

Think of it this way. If we benchmarked using only easy settings, then we send the message that StarCraft II doesn't EVER need more than a 2.5 GHz dual-core CPU, so an upgrader takes that path.

Then one day they're playing a heavy map and their system slows to a crawl. They're like, "What the hell? According to tom's all I'd ever need for this game is what I have, my dual-core 2.5 GHz CPU."

To me, benchmarks are only useful if they show a worst-case scenario. When the PC is being brutallized by AI calculations on a lot of units, that's the performance that matters because that's when you'll lag. I don't care to see benchmarks that don't showcase large battles because any PC can handle that. I want to show how bad things can get so a person knows what they need to avoid the lag.

Now granted that worst-case scenario has to be realistic otherwise it is useless, but based on a lot of Multiplayer Team-vs.-AI battles I've been a party to on the original StarCraft, there can be a lot of units on the map in the final push. I wanted to recreate that worst-case-scenario.

Certainly it's arguable that my worst-case-scenario isn't realistic, everyone is going to have their own take on that based on their style of play. But this is the reasoning behind this benchmark, and I think it remains valid.
 
Have the following setup
Phenom II x4 945, 4gb of 1600mhz ddr3, 2x 9800gtx+'s.
This system plays the game on ultra with no sluggishness.
However i am replacing the 9800's with a 5850(xfx) Tuesday so that i can sell off the 9800's in a couple of custom built computers.
 
[citation][nom]Cleeve[/nom]That was the point though. Think of it this way. If we benchmarked using only easy settings, then we send the message that StarCraft II doesn't EVER need more than a 2.5 GHz dual-core CPU, so an upgrader takes that path.Then one day they're playing a heavy map and their system slows to a crawl. They're like, "What the hell? According to tom's all I'd ever need for this game is what I have, my dual-core 2.5 GHz CPU."To me, benchmarks are only useful if they show a worst-case scenario. When the PC is being brutallized by AI calculations on a lot of units, that's the performance that matters because that's when you'll lag. I don't care to see benchmarks that don't showcase large battles because any PC can handle that. I want to show how bad things can get so a person knows what they need to avoid the lag. Now granted that worst-case scenario has to be realistic otherwise it is useless, but based on a lot of Multiplayer Team-vs.-AI battles I've been a party to on the original StarCraft, there can be a lot of units on the map in the final push. I wanted to recreate that worst-case-scenario.Certainly it's arguable that my worst-case-scenario isn't realistic, everyone is going to have their own take on that based on their style of play. But this is the reasoning behind this benchmark, and I think it remains valid.[/citation]

I wasn't criticizing your article. I was just trying to help the guy I quoted with my experience with the I5 750, since he was asking about the difference with I5 750 at 3.6 ghz. It had nothing to do with your article.
 
This article makes me hesitate in dropping $350 on the gigabyte super OC 470 i've been eyeballing. I'm tempted to get the cyclone 1gb 460 I'd prev been looking at and up to SLI a little later. Or get the 470, wait longer, and SLI that.

Any thoughts?
 
Shirosaki, must be doing something wrong. Techspot also did a review on this game. Good read, like this article. Heres a graph, where the performance of a 2.4 C2Q was much lower than Intel's new cpu's.
http://www.techspot.com/review/305-starcraft2-performance/page13.html
CPU.png


edit: Muckle, that Gigabyte with its factory o/c@700mhz is a beast. You can o/c a gtx 460 to about 820 (some come that way) and its close to a stock 470. But not one at 700mhz.
Depends how much gpu power you need, or should I say, can afford :)
 
tbh it runs fine on 1680x1050 w/ high quality settings on my q6600 with a gf 512mb 8800gt and 4gb of ddr2 ram
only thing that puts it to the test are big end-game battles in 3v3 and 4v4
 


No problem. I replied to your post specifically, but the reply was meant as a blanket for a lot of folks who feel that the benchmark is too difficult. :)
 
@Cleeve

Ive seen the benchmark but haven't ran it on my computer with fraps. I'd like to make a comment on the way the map is designed...and yes I did play it on normal speed. This is only based on my visual estimation of frame rates using ctrl + alt + f in SC2.

First, I think only the first 20 seconds of the benchmark is valid as a real "stress" because frames will constantly be increasing once the first few sets of units starts to die. Using SC2 in game frames, my old computer Pentium D 3.2 GHZ/ 3 GB DDR2 / 9800 GT goes from 5 frames to 8 for the first 20 seconds. The next 20 seconds it jumps from 8 to 14. The last 20 seconds goes from 14 to 23. Shouldn't a benchmark be able to provide consistent stress for the entire length of a benchmark. I understand that ladder SC2 battles don't last over 20 seconds but this is SC2. There are custom maps that bring my old computer down to 1 frame for much longer then 20 seconds (Marine madness, income wars, etc). Try playing a 1 hr game of marine madness on core i7.

Second, the test method seems a bit inconsistent. Wouldn't it be easier to remove human error of the test by fixing a screen at a certain location and using triggers to constantly spawn units in the screen? This way there is even distribution of stress throughout the whole test duration.

Third, I was wrong about the doodads.

I find it strange that people want to "stress test" the game using ladder games situation while I want the game tested in custom game madness situations. I appreciate this article...20 fps is still very smooth and is not abysmal performance for an RTS game. I play just fine in 2v2 ladder games at 10-20 fps. However in 3v3 games, my Pent D drops down to 5 with no battles, and 1 with battles. I still don't think this benchmark is realistically showing performance of a 4v4 ladder game.
 
Just tested it on my friends computer.

Ultra Settings 1680x1050 no AA
Athalon 2 x4 640
5770
4 GB DDR 3
no overclock

First 10 seconds was at 15 FPS. At 30 seconds, it was at 25 fps. At the end was at 40 FPS. We play custom games where he drops down to 9 fps occasionally and my Pent D would be at 1 fps.

All this is done with SC2 internal FPS and I'm just posting my results.
 


That's why I capture minimum frame rates - the minimum stays the same regardless of the length of the benchmark and the variance in the stress. The average FPS gives an indication of the playability over time, more or less.





Hasn't really been an issue, the benchmark delivers incredibly consistent results. I've found there to be less than an FPS or two average and less than a single FPS min. If you run the bench three times you will get a very consistent result, remember it's CPU limited so the actual window of view has little effect on the frame rate. It's the calculations in the background that are slowing it down, where you look doesn't seem to matter that much.




Well, the trick is making a bench that everyone can find useful to some degree and that's what I've tried to do here. Even though this isn't a perfect bench for everyone I think it delivers useful info and gives folks an idea of what to expect. Take into account your own particular scenario's variables to taste. :)
 
No one with a 5870 play in medium. CPU should also be done in high or ultra. That's a real scenario.

pjladyfox
Athlon II X4 640 is a well around cpu for any given game for the price unless you're for an HTPC AII X2 lost their sweet.

 
Could you post the Tom's Hardware Benchmark to EU.BATTLE.NET as well, please
 
while the highest graphic quality of the game is really good,
playing the game with low settings doen't remove anything.
the game stays the same.

i have a core 2 duo E2200 overclocked to 2.93ghz ( 1.33 of it's original clock or 25% improvement ) 4gb ddr2 800 mhz ram and a good old geforce 8800gts with like 380 mb or something ... with windows 7 64-bit. I got occasionnal lag while playing custom maps with mass units so i used minimum settings from then ( 1920X1080 but without any effects )
 


You don't understand the test. Testing on high or ultra potentially moves the bottleneck to the graphics card. The purpose of the CPU benchmarks is to test the CPU bottleneck, not the graphics card ability.

In any case, our other tests suggest that the CPU bottleneck is so strong that High or Ultra graphics settings would make little difference, as long as your graphics card is a GTX 460/5830 or better.
 
[citation][nom]madass[/nom]Are you guys sure the NV cards didnt beat the radeons due to bigger frame buffer?[/citation]

i'd be more concerned about nvidia optimazations being what gave nv the win here, seriously some nvidia optimized games run like crap on ati's , best example is nwn 2 , i actually went up from a GF9800 GT to a radeon 5770 and lost 10 frames on that game , (every other game runs beter on the 5770) gratned the game is still very playable at amx settings , but that's beside the point the game was heavily optimised for nvidia , and it runs worse on better ati cards because of it . this coudleasily be an optimazation issue here.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom