• Announcements

    • HEAVY265

      Attention Soldiers Operation Fury Needs you!   02/20/2020

      Attention All Soldiers, Operation Fury needs you.  You need to choose a side and sign up.  
      For more intel on Operation Fury Please click HERE Please go to Special Event Forum (here), And sign up for allied or axis.
      This will be a CRS Lead event on both sides.  Xoom will be heading up the axis side and Heavy265 will be heading up the Allied side. This will be for bragging rights.
      Why are we asking players to sign up you ask. We are trying for a role play experience.   We want this to be a true realistic event.  
      So get up and sign up and let's make this the best event ever!!!!!!!!!!
      Give me your war cry, grrrrrrrrrrrrr
      Heavy265 **out**
Sign in to follow this  
Followers 0
AHWULF

1.31 open beta (soon)

39 posts in this topic

Most likely:

Intel Processor (core2 duo+)

OSX 10.5 and 10.6

2GB of RAM (beta may require 3 unless I figure out the #$@#$ bug I'm working on today).

Video card (no integrated video) with 256MB.

The biggest question is still what video card constitutes bare minimum. The beta will help but ultimately it may depend on what Apple can tell me. My GT120 is crappy but so far works OK even with the current version. The "M" cards in laptops and older iMacs are the biggest concern.

Your best bet is a decent iMac with an ATI 4xxx card. It's the lower end iMacs / MBPs that are fuzzy atm.

If it won't play with my iMac (core2 duo+), I won't be getting a new computer just to play this game... I've already been down that path.

The only other option will be to put my finger on the button unsub_button.gif

Share this post


Link to post
Share on other sites
If it won't play with my iMac (core2 duo+), I won't be getting a new computer just to play this game... I've already been down that path.

The only other option will be to put my finger on the button unsub_button.gif

Same here, I just hope my Geforce 9400M can keep up.

Share this post


Link to post
Share on other sites

BLOO tells me on the PC side they don't support the M cards. I'm not willing to say that until I see it not work (and hear from Apple).

Share this post


Link to post
Share on other sites
Read Gophur's specs again Fundin. OS X with Mac specs (same hardware specs) is included there. :)

Oh yeah thanks, well hidden to the scanning eye :)

Share this post


Link to post
Share on other sites

Motormouth made me write something up for posting on the website. Mostly the same stuff as above. Not sure when it appears or where.

Share this post


Link to post
Share on other sites

Ok, thanks for explanation.

By the way, would 1.31 benefit from SSE2/SSE3 optimization, or Mac version would also be left "out of the room" because noone bothered to turn on one compiler checkbox?

And regarding this:

Couple that with Apple's crummy opengl drivers and running on the exact same hardware' date=' the Mac version can often be 1/2 the FPS of the PC on lower end hardware [/quote']

If Apple do release drivers for bootcamp, who could this happens that Apple's driver for Windows are much more optimized then for Mac OS X? As far as i understand, currently bootcamp/windows version works much faster then Mac OS X version on the same hardware.

Share this post


Link to post
Share on other sites

I have a feeling Apple uses the video card manufacturers's drivers under Windows and just tests them.

I've tried SSEx compiler option but didn't see any improvement.

Share this post


Link to post
Share on other sites
I've tried SSEx compiler option but didn't see any improvement.

Well, i was kinda kidding about "single checkbox on".

As a developer who used to work a lot in the SIMD optimization, i would say that no compiler would be able to generate descent SIMD optimized code for SSE2/3/4 comparable with manually "fine-tuned" intrinsics.

What is done usually - developers, using profiler, isolates pieces of code which are highly loaded, and rewrite em to SIMD. Then use SSE-enabled code under some compiler flag/runtime flag, leaving "reference" implementation (plain c/c++) as a fallback solution in case if compiler doesnt support such optimization or runtime option turned it off.

So it is not the compiler option actually, its a matter of programming.

Share this post


Link to post
Share on other sites
Ok, thanks for explanation.

By the way, would 1.31 benefit from SSE2/SSE3 optimization, or Mac version would also be left "out of the room" because noone bothered to turn on one compiler checkbox?

And regarding this:

If Apple do release drivers for bootcamp, who could this happens that Apple's driver for Windows are much more optimized then for Mac OS X? As far as i understand, currently bootcamp/windows version works much faster then Mac OS X version on the same hardware.

You can't believe the trickery Apple does on boot camp (even with snow Leopard) which I witnessed. Installing OLD nvidia driver, OLD realtek driver, not enabling AHCI (thus NCQ) by not doing couple of byte changes in MBR (which Windows demands). If you test/play game on Windows, make sure you do update drivers your way while MBR trick is real complex and won't work with Nvidia chipset. I speak about 20MB/sec MAX on a decent SATA2 chipset/new hard disk. So, Mac Mini ends up in "bottom 10%" on pcpitstop.com

I suspect they are horrified that their customers may switch to dark side because it performs, no other explanation than that.

ps: As you know, realtek is a soft audio chip and has high end specs (24bit/96khz) so if there is a bug in old driver, it will choke the OS/game whenever it makes some sound.

Share this post


Link to post
Share on other sites
Well, i was kinda kidding about "single checkbox on".

As a developer who used to work a lot in the SIMD optimization, i would say that no compiler would be able to generate descent SIMD optimized code for SSE2/3/4 comparable with manually "fine-tuned" intrinsics.

What is done usually - developers, using profiler, isolates pieces of code which are highly loaded, and rewrite em to SIMD. Then use SSE-enabled code under some compiler flag/runtime flag, leaving "reference" implementation (plain c/c++) as a fallback solution in case if compiler doesnt support such optimization or runtime option turned it off.

So it is not the compiler option actually, its a matter of programming.

Well, yeah, I was sort of joking too :-)

My biggest issue about using this is making sure which Intel chips APple used support which of the (4) variations. There are SSE instructions in various places but mostly PC libraries have them. The PC is compiled with the flags but I'm not sure there is much code difference either in the main application.

Right now the biggest performance killer is in the app->opengl framework->driver area which is what I need Apple's help on. This is where the PC/Mac versions differ most, and there is no easy way to know what we're doing wrong without someone who can look at the innards of that area. I know the apple gl extensions can help but there isn't enough documentation to know how to measure the problem.

Virtually all recent Mac models have more than enough CPU (core2duo+) for the code base, its that connection between the CPU and GPU that is the problem. So SSEi won't help much until we can unstick this area.

Share this post


Link to post
Share on other sites

It won't be this week. No one is in the office. Even my net connection to the office is down for some reason. Still have plenty I can do at home.

Share this post


Link to post
Share on other sites
????????????????????????

Did you ignore ahwulf's post? There is no one in the office till Monday.. He will find out what happened then.

Share this post


Link to post
Share on other sites
Sign in to follow this  
Followers 0

  • Recently Browsing   0 members

    No registered users viewing this page.