Half-Life 2 Upgrade Analysis

RaidRaider writes "Hardware Analysis makes a good effort at answering the question that has been on every enthusiast's mind; what are the specs for a baseline system I need to enjoy Half-Life 2 to the fullest? They take a good hard look at the Steam survey results and work their way up from there, replacing CPUs, graphic cards and add more memory. They back it all up by offering the demos used up for download so you yourself can gauge exactly what kind of upgrade you need."

Comments

Works fine for me...

[omeglidan]

Will upgrade when Duke Nukem Forever comes

Don't forget Gamespot

[f4llenang3l]

Gamespot had a comprehensive article in which they did thorough bottlneck benchmark testing to determine what you should replace to get the most bang-for-your-buck for HL2 out of a hardware upgrade. They cover everything, from different DirectX modes, to onboard sound vs. sound card, and of course processors, RAM, and video cards. It's a great way to figure out how to best spend a hundred bucks to make HL2 a much better game for you.

Don't Forget

[Mr.Dippy]

You also need an internet connection. If you want to install the game on a box that doesn't have a connection to the internet you are screwed.

Re:Don't Forget

[10537]

And also don't forget that if you bought the retail boxed version, not only must you have an internet connection to authorise/decrypt the files, but you still need the disk in the frickin drive to play. And woe betide anyone who uses a no-CD patch, for you shall be cast in to the fiery pits of hell...

Re:Don't Forget

[Sentry21]

I will admit, I downloaded the game off the internet. Here are a few things I noticed as a result:

1. The game itself runs amazingly well on my system - the recommended settings not only look better, but also play much smoother, than Halo PC on minimum detail at 640x480. I never would have purchased the game because of the assumption that my system would almost, but not quite, be able to handle it. Now I am thinking otherwise.

2. The physics engine is amazing. If this engine will run on my friends' PCs, and if Counterstrike: Source is as open an environment as Half-Life 2 is, then we will definitely all buy a license to it for use as our latest LAN game. I never would have considered this without playing it first.

3. The gameplay is very well done, except for the three hours I spent in that goddamn airboat thing. It's freaky.

4. Before playing the game, I was very doubtful, even after all the good things I'd heard. Now, I am raving about all the cool things to my friends, who DO want to buy the game. I've probably made a few sales already, and will likely add myself to the list.

Finally, here are my specs:

Dell Inspiron 5150 Laptop
P4 2.8HT
512 MB / 60G 5200RPM HD
Nvidia GeForce Go FX5200 64mb

As I said, this system will not run Halo smoothly (I blame Halo - everything else runs fine), but Half-Life 2 was not only smooth, but beautiful as well. I can't imagine what it would look like at full detail at 60fps. Because it is capable of this, I will likely buy a copy soon, if only to show my support for a game so well-coded.

DVD Statistics...

[Spuffin]

I would like to point out that the DVD statistics may be a bit skewed. I took the hardware survey twice and both times it detected I had a DVD drive which is not the case. Apparently the scanner detects the Daemon tools drive and concludes you have a DVD drive. It also fails to detect that you also have a CD drive. Not that I'm against putting more games out on DVD, I just think these statistics should be taken with a grain of salt.

Undetermined 80,081 6.23 %
CD-Rom 194,344 15.12 %
DVD 1,011,326 78.66 %

Average Sucks?

[th3walrus]

I find it disturbing that performance on the average system (according to Steam stats) is considered unacceptable.

This was a rather lacking article.

[TellarHK]

After reading through the linked pages, I really didn't come away thinking I'd learned anything of value from this site other than the fact the CPU bottleneck is pretty large. The comparison really lacked in depth and breadth, as they only appear to have used two video cards, 4 processors and a couple variations on RAM. Also, the qualification for "acceptable" seemed a little arbitrary as it wasn't the 30 or 60 fps I usually see used as a rating of acceptability. (100 fps for the psychotic benchmarkers)

Also, the structure of the article could definitely have used some help. There was no clear delineation between the various tests, such as each page being used to display one processor with multiple card benchmarks. The information was largely there, but it definitely wasn't all that clearly stated. Maybe this was due to the extremely limited number of cards they were able to test with, but it still leaves me wanting to know more.

My own experience with Half-Life 2 would lead me to saying that running it is no problem, and that running it with a 2Ghz (core clock) processor and a Radeon 9800 Pro in 1024x768 should be just fine. Just keep the textures at Medium and it'd be all set. High detail textures seem to be the system killer on machines such as those.

Well duh

[SmallFurryCreature]

The daemon tool is a dvd drive just a virtual one. If it was easily detectable that it was not a real drive then it wouldn't be half as usefull.

Off-topic/video card prices

[dackroyd]

What the frick is happening to the prices of video cards?

In the old days (ie three years ago) the price of each card would fall over time, and then fall dramatically once a couple of video card 'generations' had passed.

Nowadays it seems that the price of existing video cards is stable and that newer cards are coming out at ever increasing prices.

For example the 9800 pro came out over a year ago but is still $350-400 dollars. The new ATI cards coming out are over $500...

This can't really represent the mass market for video cards can it ?

Thanks!

[M.C. Hampster]

I think most surprising for me was how having a dedicated sound card, rather than using on-board sound really affects the performance. I've never been one to care for sound quality, so I haven't bothered with purchasing an actual sound card for my system, but I didn't realize that the onboard sound uses up CPU.

That's probably the only upgrade I'll really need to make before I get the game.

Authenticity?

[jm92956n]

I didn't take the survey (as I'm on a Celeron 500 with integrated video that's entirely incapable of playing games), but I'm curious to know if the detection tool allowed users to correct mistakes or otherwise alter the data. I wonder about the:

- 1,500 people still stuck on 14.4 modems.
- The 94 people who attempted to play the game with 32 megabytes of RAM or less.
- The 111 people with processors no faster than 200 mhz.
- The one person with a 4-CPU system!
- Lastly, both of the following screen widths had exactly one user: 5 pixels and 3,072 pixels.

Were people with these systems actually playing Half-Life 2, or can it be better explained as users with a sense of humor?

My system

[Nosferatu Alucard]

I skimmed the article, and I noticed they didn't show any graphs from Nvidia based cards, just ATi. I also read it where they said that they couldn't get their hands on the Nvidia cards in time for the test. What im wondering is why they end up with your only option is an expensive card and a Socket 754 processor as the "best" option. I'm running HL2 just fine, I only got lag when the map loads ended, and from then on I never saw a bit of skipping. I'm running on a +2800 AMD Barton with a 6800 and 512MB RAM. I get 50-70fps with 4x Anti Aliasing, 8x filtering and all my settings at either high or medium. I don't think I'd consider this a fair article. TO quote them "An AMD Sempron 3100+, paired with a ATI X800 Pro videocard and a nForce3-250 motherboard offers all the performance you need to get started." My chip and video card cost over $100 less than their setup and it runs the game beautifully. I'd advise not paying much attention to this article.