More pages: 1
... 34 35 36 37 38
39 40 41 42 43 44
Thursday, August 7, 2003 | Permalink
I have returned to civilization after an over two months long holiday spent with my family. The weather has been wonderful and I have had a lot of fun.
Had to do wome work too though, been working pretty much all summer on the report on my work at ATi for the university. I'm going to present it early next month. I've also spent some time writing a new framework. I have fixed all the problems with earlier frameworks, though there are still some work left to be done.
I've of course snapped a zillion pictures too. Will put up some galleries soon.
[ 1 comments
| Last comment by davepermen (2003-08-08 09:01:10)
Sunday, June 8, 2003 | Permalink
I'm going home to my family over the summer. Will be gone for a couple of months or so, don't know exactly. But I'll be coding demos as usual, though I may not get the chance to upload it all that often since I won't have any internet access.
Anyway, have a nice summer and have a lot of fun. I will.
[ 4 comments
| Last comment by Sterlinger (2003-07-02 08:58:38)
Thursday, June 5, 2003 | Permalink
I have added some benchmarking tools into my framework. All demos compiled with it from now on will have built-in benchmarking support. Additional information here
As of the time of this writing only the VolumetricLightingII demo has been updated with benchmark support. I will update more demos later on. If there's any particular demo you'd love to have benchmarking support in, just post a comment on this news item.
Why am I doing this? Two reasons. The first is that there has been a lot of talk about cheating in drivers lately. The sheer amount of demos I keep outputting here would make it unfeasible to cheat through shader replacement. Given that anyone can record their own flyby paths it will be impossible to use tricks like static clip planes, selectively omit framebuffer clears or anything of that kind.
The other reason is of course that it will actually provide us with some interesting performance characteristics data for the latest and greatest features on our cards. Another tool in the reviewer's toolkit.
These demos now has been updated with benchmark support:
[ 5 comments
| Last comment by davepermen (2003-06-07 00:00:05)
Tuesday, June 3, 2003 | Permalink
How to build a [H]ardocp equivalent site in a few simple steps
First we must choose a Linux distro. Since we think we're so much smarter and more illuminated than the masses we'll use Gentoo. Or as we prefer to call it, [G]entoo. That didn't make sense,
well neither do we, so let's go on with the installation.
We'll start with a clean machine and boot up from the gentoo CD. Once the kernel has loaded we must ensure that our network is up and running. On the console, type the following:
#> ping www.nvidia.com
Note, it's of highest importance that we always ping nVidia before going ahead with anything. If you're not getting any response, try adding this flag:
#> ping -fanboy-rhetoric www.nvidia.com
Once our network is running we must put a bandwidth limit on our network connection. This is the give the false impression that our site has visitors and the server is bogged down. This will improve our chances of receiving boards to review. This can either be done by configuring our system, but in true [H] spirit we'll do it the [H]ard way. We'll just tie a really strong granny knot on our network cable.
Since we're über-leet we're going to use a stage1 installation. This means we'll compile everything ourself. To ensure optimum performance of the final system we must set the proper optimization flags. Edit the /etc/make.conf file.
#> nano -w /etc/make.conf
In there you'll find some generic optimizations such as:
These are safe optimisations, but hardly generate the best performance. Add the following flags:
-mgpu=GFFX -march=NV3x -static-clip-planes -omit-frame-buffer-clear -shader-replacement
These optimizations are not safe though, but valid and they are definitely not cheats dammit!!!
Now we can go ahead and install our system:
#> emerge system
This will take a long time, typically several hours, so that will leave you with plenty of time to come up with some random nonsense to post on the frontpage as soon as we're done.
Alright, once that has finished we want to install some basic services. First we need to install our database.
#> emerge mysql
Now we want our newsediting service and of course a forum.
#> emerge news
#> emerge forum
At this time we need to choose a Southpark character. Kyle is already taken though, but there are plenty of choices that will suit us just as well. A fat noisy egomanic would work, so Cartman would be fine. Or we may choose Kenny, a guy with a narrow view mumbling some random nonsense. Both these would fit our [H]-compatible site.
Alright, so the tasks above has completed and you have your news service and forum installed. Now we need to setup a site policy. Typically a site owner would have to figure out a policy of his own, but we'll use a shortcut and opt for an establish policy:
#> scp www.nvidia.com:8080/.policy ./policy
Notice that we're using secure copy, this is to ensure that nobody notices we just grabbed our policy elsewhere.
At this time we would want to send a 16-digit creditcard number to our favourite IHV hoping to maybe receive a "package" or two. Chances are that they will just empty it and leave you in the blue, but we're just clueless anyway and believe capitalistic principles such as using your market position to get advantages should only be allowed for certain small and friendly companies like nVidia and not by mega-corporation crooks like Futuremark who tries to force vendors off the market by using pure force. So let's just go ahead:
#> scp ./creditcard_number.txt www.nvidia.com:8080/creditcard_number.txt
Right, so our site should be up and running at this point. Now we need to hire a partner to offload some of the work of filling the site with clueless fanboy rambles. Note that only site owners are eligible for a Southpark character. Co-workers can only get normal names like Steve. As long as you find someone who's clueless, holds strong opinions based on info received from the Iraqi information minister, responds on feedback with flames, thinks he's sophisticated because he can look up information in google, can't take constructive criticism, thinks he's cool because he has the power to edit the pointed out errors on his site with even bigger errors, questions other peoples credibility when proven wrong, makes up his own rules on what it means to be a developer, can't take losing an argument, can't discern sarcasm from statements, thinks he's cool because he can threaten to sign up people for spam and thinks that forwarding a conversation including all the above attributes to university admins will achieve anything else than looking stupid; if you find someone who meets these requirements, and are happy with a standard name like Steve, then you've found a perfect partner for your site. Being ghei like the original is a plus.
At this time you should be pretty much set. Enjoy your [H] equivalent fanboy rambling site.
[ 25 comments
| Last comment by Iraqi information minister (2003-08-09 20:56:40)
Sunday, June 1, 2003 | Permalink
Time for another volumetric lighting demo.
Some nice volumetric lights to look into until you've burned the macula.
The map is the DM-Tutorial map from UT.
[ 5 comments
| Last comment by Humus (2003-06-06 16:13:57)
Friday, May 30, 2003 | Permalink
I have updated the Mandelbrot and MandelbrotSet demos with a zoom console command to let you define an exact zoom for comparion purposes.
[ 1 comments
| Last comment by tobias (2007-03-28 10:51:01)
Wednesday, May 28, 2003 | Permalink
Another demo to waste your time on staring at in vain.
This time it's a nice electro-flash procedural texture with a only 9 instructions long fragment shader.
[ 1 comments
| Last comment by davepermen (2003-06-03 08:57:16)
The state of ethics in 3D graphics
Friday, May 23, 2003 | Permalink
We live in sad times. Around the time of the release of Radeon 8500 news spread on the web the ATi applied cheats
to gain performance in Quake3. SiS has been busted
occasionally for lowering IQ for performance in benchmarks. Not too long ago Trident was accused of cheating
too. Recently nVidia officially joined the crew when their cheats
was exposed. This time it went as far as to Futuremark having to issue a statement
and provide a patch
to disable the cheats and a detailed report
on the cheats applied. No less than 8 different cheats has been applied by the driver which added no less than 24% more performance. This puts the NV35 into a less pleasing light. Instead of being the chip that brings nVidia's floating point performance up on par with ATi the painful truth again after disabling the cheats looks to be that the floating point performance still is crap. At this time I also have no hope that any NV3x well ever perform well in that area, the hardware is just that slow, as patched 3dmarks proves and other applications has proven for some time. It should also be said that Futuremark found oddities in one test where ATi seems to get a small but suspicious performance drop with the patch applied. The jury is still out on that one though, but in all likelyhood the truth may be less beautiful than one might hope.
It's sad. It's really sad.
But as usual, there's something to learn. The lesson is NOT that 3dmarks is useless as a benchmark that some sites, which I refuse to name and link to, has implied. And it's absolutely not that one's cheating justifies another one's as endless fanboys are ranting. One may have have objective opinions on 3dmarks, but the real lesson is that reviewers need to spend more time looking at Image Quality and look for rendering anomalities. Reviewers need to break away from traditional benchmarks and look for alternatives. Benchmark games that are less known. Write your own benchmarking tools. I have little hope on the last one though given the enourmous amount of ignorance out there. Most review sites can hardly discern a pipeline from a pixel. The last, and maybe the most important point though is that we as consumers need to be aware of this. We need to demand that this does not happend. We all vote with our wallets. We much make it so that it's not worth it to cheat.
[ 9 comments
| Last comment by Doomtrooper (2003-05-28 01:22:56)
More pages: 1
... 34 35 36 37 38
39 40 41 42 43 44