Half Life 2 Performance Evaluation

www.bit-tech.net/article/148

 

cheers.

 

Damn good article. Another job well done, biggles

*n

 

Phew, I should be alright to play it then . Nice article.

 

you know the more i read articles like this and see screenshots of the top cards i realise that my 9600xt is no longer as good as it once was a year ago. At least i finally got the game though and made use of the voucher.

i noticed you mentioned the "blips" and i to experianced those around saving. Valve aparantly released an update for steam to fix this but ive not played it since i finished it on the 17th. (yes i was amongst the sad lot who set his alarm for 8am on the 16th )

Good article though, very different and refreshing to most of the others.

 

I was aware of that update, but I still experience these blips strangely enough, despite those, frame rates are completely playable.

 

my x800 plays hl2 great, specially now ive opened up all the pipes turning it into a xtpe instead of a pro (thanks kammy)

but still, nice article biggie

certainly getting a flow for these now aint you?

EDIT: havent i seen that hl2 logo before?

 

That's 110% ours. Unless you mean from valve.

 

of course i mean from valve, they are the only people who have it copywrited yes?

 

Great article! what ever happend to the comparison of images tho?

I just had a quick compare of the airboat and the only differnce i can see is the frame around the seat looks a little alaised on the Nvidia one which is a minor detail at best.

Edit: on the second image, the color for the x800XT seems alot darker, almost as though the scene is lit differently (see airboat handelbars)

I also get the idea though, that you didn't include comparisons becuase both cards render this game very well, to the point that it's hard to find any differences.

 

I agree, this article is good, but there must be some bypass of common sense at Bit-Tech. Let me explain; im a daily visitor the Bit-Tech.net, and i read almost everything put on here. It just seems incredibly stupid to do all your benchmarks on such high end systems. I imagine that i am not the only one that feels these benchmarks are almost completely useless as they have to correlation to their own systems.
You even say yourselves "Most of the title is very CPU limited", lets be perfectly honest with ourselves, what is the actual number of people who read that article who have, or even know someone who has an AMD64 FX-55? Understandably the test is to prove the best card, but how can that give anyone with a reasonable system any ideas to what the game will perform like on their own machine?
Im not trying to have a dig though, your tests are very thorough and are respected across the internet. Maybe just do the tests on a CPU that some people can relate to.

 

The game does render very well on all cards, which is why I've tried not to be overly nit picky about image quality - the differences that are there aren't necessarily bad in any stretch of the imagination. The only bad thing that I found with regard to image quality is the poor distance clipping that I can't seem to find a hack for at the moment.

Arena_08, thanks for signing up to comment, I feel that feedback is important in this sort of article/review style, because we are trying to relate to the common user as much as possible.


Would you feel that things would be easier to understand if we considered our high-end system to be a 4000+ (2.4GHz), while mainstream sits at 2.0GHz (circa 3200/3400+)? I'll stick with the FX-55, but I am open to opinions as to how far you feel we should downclock the CPU to represent a more realistic gaming experience.

I'm after your guy's opinions here, this is only the 5th time that we have used this style of review, and the feedback so far is very good - we just need to fine tweak things a little to give a more realistic set of numbers to what we can expect people to achieve. The problem is that if we reduce the CPU speed too much on the high-end video cards, we do not see the differences between a Radeon X800 XT Platinum Edition and a GeForce 6800Ultra, as ultimately we are really beginning to limit performance to the capabilities of the CPU, rather than the video card.

 

i'm guessing a mainstream setup would be around the 3200+ mark; most people i know either own a CPU around that level, over overclock to that level.

 

I don't see what your point is, most games today arn't very CPU limited and most of the people that are going to be reading the reviews or are serious gamers will have at the very least a 2.4B P4/ 2500+ Athlon XP. Furthermore anyone who buys an x800XT PE will most likly have a CPU every bit as expensive (p4EE, FX53/55). Even 9800 pro buyers will most likly have a system specs up there as i dosn't cost much now for a 3Ghz p4 or an AMD 64 3000+

Also: Bit-tech is a tech site, most if not all the people here are interested in the bleeding edge of technology, not how some game performs on 2 year old hardware.

 

We're already at 3200+/3400+ with an FX-55 at 2.0GHz if my calculations are correct.

2.2GHz with 512K cache = 3500+
2.0GHz with 512K cache = 3200+
2.0GHz with 1MB cache (what we are using for 6600GT/9800 Pro) = ~3300-3400+ (I think)

We are interested in bleeding edge tech, but, we want to be as realistic as possible with our real-world gaming tests. After all, we are trying to be as realistic as possible without slowing things down to such a level where the CPU takes over and controls how well the video card performs - it's just finding that right balance between speed and realism. I do think that Arena_08 has a point, but I am interested in what others think too - you guys read the site, I just write about my findings!

Maybe the 6800Ultra/X800 XT PE processor speed is alright, but what about for the 6800GT/X800 Pro? Should we be running something more respective like a 2.4GHz, 1MB cache A64 as our baseline CPU? I'm all open for ideas!

 

i agree with you using your fx-55 with the x800's and 6800U's due to the reasons you said in the above posts, and when comparing these to lesser cards i.e. you have to keep the same processor which is great for gfx card reviews.

but when your doing HL2 performance it would be nice to see how it runs on an average sytem i.e. a 9800 pro etc, as this is what most people have.

http://www.steampowered.com/status/survey.html

the people who use steam have this hardware. ^^ youve probaly seen it.

looks like the 9800 pro vs 6600GT is a very close battle thou.

bring back the overclocked mobile barton

i dont know how possible this is, how about comparing all 4 cards on the same chart, hi-end vs mainsteam. basicly users who are considering if to go the whole hog and get a 6800 over a 6600GT is worth is for example.

 

Thanks Bigz, the test at 2.0GHz is much appreciated. I've got an XP2100 and recently got a 6600GT and my video settings and framerates are similar (I don't use AA or AF with HL2 but it's mostly very playable).

Now, on to that upgrade...

 

yes, a 2.4B/2500+ is about at that point. A few people still have <= those processors, but most of them have overclocked their processor to acceptable levels or are thinking of an upgrade pretty soon and want to know what to upgrade to, not what their upgradeing from after all, if they have the hardware, they can run the tests themselves, right?

 

I know what Bit-tech is about, and I am a serious gamer. Its just some of us cannot afford to spend 2 months wages on a CPU, its just not a realistic possiblity. Im just trying to say, with such a high level CPU its difficult to relate to your own PC.

 

in which case you can't afford much better then a gf fx 5200 either. No one says you have to spend a grand on a processor, a 2800+ AMD 64 is all you need and you get get it for a price thats a few days wages if you work at Mcdonalds, not 2 months! Out of curosity, what is your CPU?