gulogo.gif  
 
1. Hiatus
2. RIP, Satoru Iwata
3. Let there be Robot Battles
4. Regarding pixel art!
5. 16-bit Star Wars
6. Goodbye, Spock.
7. James Randi Retires
8. More Star Wars on GOG
9. Archive.org gives you DOS Games
10. Ralph Baer, RIP.
1. Quickie: Impressions June 2014
2. Quickie: Penny Arcade Episode 3
3. Quickie: The Amazing Spider-Man
4. Quickie: Transformers: Fall of Cybertron
5. Quickie: Prototype 2
6. Quickie: Microsoft Kinect
7. Quickie: X-Men Destiny
8. Spider-Man: Edge of Time
9. Quickie: Transformers Dark of the Moon
10. Quickie: Borderlands GOTY
1. Musings 45: Penny Arcade and The Gripping Hand
2. Movie Review: Pacific Rim
3. Movie Review: Wreck-It Ralph
4. Glide Wrapper Repository
5. Movie Review: Winnie The Pooh
6. Musings 44: PC Gaming? Maybe it's on Life Support
7. Video Games Live 2009
8. Movie Review: District 9
9. Musings: Stardock, DRM, and Gamers' Rights
10. Musings: How DRM Hurts PC Gaming
Main Menu

Affiliates
X-bit labs
The Tech Zone
Twin Galaxies

Login






 Log in Problems?
 New User? Sign Up!


 3D Acceleration History and Glide Wrapper Comparison @ GU
 June 16th, 1999 by Chris "Stories" Kim

[History of 3D Acceleration] - [Purpose of Glide Wrappers] - [Historical Wrappers] - [Current Wrappers] - [Comparrison of the Current Wrappers] - [Conclusion]

History of 3D Acceleration: Most of you guys are already readily familiar with Glide Wrappers because Glide Underground originated as a Glide Wrapping community.  Just incase you don't know about Glide Wrappers, here is brief description about Glide Wrappers.  Prior to 1996, there was not a standard for 3D graphics.  Everything was just software rendered graphics, and therefore very pixelated as there was no dedicated graphics rendering hardware.  But that all changed in 1996 with the release of the Voodoo Graphics chipset from 3dfx.  This new technology known as 3D Acceleration was very new, therefore very expensive.  At the time most Voodoo graphics boards costed well into the high $200 range and was built for the faster PC's.  Another disadvantage of the Voodoo graphics was that it as a PCI slot, 3D only gaming device and therefore required a seperate videocard for 2D graphics such as Windows.  Along with the Voodoo graphics boards, came a 3D Gaming API known as Glide.  Later, many other graphics companies followed with attempts at 3D gaming boards that combined 3D Acceleration and 2D Acceleration into one slot.  Most cards couldn't even come close to the sheer speed and quality of the Voodoo graphics board.  In late 1997, NVidia released their Riva 128 graphics chip, it was as fast as the Voodoo graphics chipset, however did not have very pretty image quality.  In early 1998 3dfx released their Voodoo2 graphics board, a huge successor to the original Voodoo graphics chipset, however it was still a single 3D only add-on board.  3dfx would hold onto this 3D graphics reign for a long time.  Many other companies released many other boards along that year, none of them even competing with the Voodoo2.  Most graphics boards were on par with the original Voodoo.  Until NVidia souped up something really actually rivaled the Voodoo2, the NVidia Riva TNT board.  Released mid 1998 the TNT was a huge success in terms of marketing and in product quality.  It had the speed of the Voodoo2, with better 3D quality, and a top of the line 2D graphics engine all in one slot (AGP or PCI).  However, all the older games prior to the major incorporation of Direct3D 5.0 and OpenGL, Glide dominated the supported 3D API.  3dfx held onto this API and would not let other companies build drivers for Glide, this is where Glide wrappers come in...

Purpose of Glide Wrappers/What They Do: What is the purpose of a Glide wrapper you may ask?  If you are an avid gamer you will probably know that most modern games have 3D Acceleration with support for one or more of three major 3D gaming API's.  These include Direct3D, OpenGL, and Glide.  Direct3D is an API by Microsoft which is the most widespread and used.  Most graphics cards come with a Direct3D compliant driver.  The only draw back to Direct3D, however, is the fact that the drivers will vary from chipset to chipset having different types of compatibility for features found on a card.  OpenGL is another widely used API mainly for 3D shooters such as Quake II, Half-Life, and a other large variety of games.  All new cards have a OpenGL driver, but many older cards either don't have an OpenGL driver or don't have very good drivers.  The least used and oldest gaming 3D API is Glide, a graphics library only for graphics cards made by 3dfx.  This API was very easy to program in and was the longest lasting API till today.  Today, there is no room for an API like Glide as most games nowadays come with either Direct3D or OpenGL out of the box (i.e. Descent 3 has native support for all three API's).  But many older games only had support for Glide, and since 3dfx holds onto Glide, they will not allow any other companies to make drivers for it because having Glide support on their cards is one of the main reasons many people will buy 3dfx cards.  But, because native drivers won't be possible for graphics card companies, little people and sometimes even big companies, will make a "wrapper" to translate Glide graphic calls into a more familiar language for most graphics cards such as Direct3D.  So, in a sense, what Glide wrappers do are let those people without a 3dfx based videocard to play Glide only, or 3dfx only based games with 3D Acceleration enabled.  Most games aimed at only use the Glide 2.x so that is what most wrappers were developed for.

Historical Wrappers: Many wrappers have come and gone, some being better than others.  The whole wrapping scene started out with Let, and his wrapper called GliD3D.  It was an attempt to wrap Glide into Direct3D calls.  He was wrapping both Glide 2.x and Glide 3.x.  However, a few months into his development he received a nice little letter from 3dfx to take down the "offensive" material which was an infringment on the Glide 2.x Software Developers Kit (SDK) End User Licence Agreement (EULA) that he had used the SDK for illegal purposes other than development on a 3dfx product.  The wrapper wasn't too far in progress when it was stopped, it could play Need for Speed II SE to some extent with 3D Acceleration enabled, but with horrible graphic corruption.  Then along came many other wrappers, most not too different from others.  Then along came XGlide, one of the most promising wrappers that was quite decent but only played UltraHLE.  UltraHLE at the time was the craze because it was a huge impact on both the Emulation and Glide Wrapping scenes.  UltraHLE was the very first N64 playable emulator, however it was only accessable to those who had a 3dfx based graphic card.  Thus started the Glide Wrapper craze for those people with other graphics cards other than a 3dfx based product.  XGlide was very popular, but was very unoptimized and thus very slow.  The graphic quality was decent, but there were some misplaced polygons and lots of redraw all over the place.  Helder (author of XGlide) finally lost interest in XGlide and it was dropped, but at the same time another new wrapper XGL200 was in development.  This wrapper was created by Scott Cuttler and is still in development today, however at the current time it was developed with the Glide 2.x SDK.  At the time XGL200 was still in it's infancy and lacked many features of the Glide language and was only playable on UltraHLE (which used relatively few Glide calls).  The wrapper progressed a lot, and was updated almost everyday.  But then 3dfx soon saw what was happening with the efforts of XGL200 and eventually halted it.  A wrapper also called Clide was another wrapper, but unlike others was developed without the SDK.  It was very fast, but lacked a lot of the image quality that other wrappers had, a side effect of development without the SDK.  The wrapper gained a little popularity, but soon died out because of lack of interest in the project and time.  The wrapper was not taken down by 3dfx at the time when 3dfx cracked down on a large population of the wrappers available on Glide Underground.  MGlide another wrapper created by Nanotech was in progress for a long time and was developed without the Glide 2.x SDK.  It was quite advanced, but did not concentrate on UltraHLE at all, Nanotech was more focused on playing the real reason of Glide Wrappers, GLIDE ONLY GAMES!  And this is where we stand today, only MGlide and XGL200 are left.

Current Day Wrappers: Today, there are not nearly as many Glide wrappers available due to the fact that 3dfx will absolutely not tolerate development with the Glide 2.x SDK.  Wrappers today must be developed using 3rd party information available freely through the net without any SDK commands.  Since development must continue without any direct information from 3dfx, development has slowed down considerably.  Scott no longer updates his wrapper daily, we are lucky if he updates every other week.  MGlide is developed at the same pace as usual, and continues to progress more and more.  The UltraHLE craze has died out and the main purpose of Glide Wrappers is not longer to play UltraHLE, but rather to play Glide only games as the original intention was.  Both wrappers have their own strong points and are very good wrappers with pretty decent compatibility and graphic quality.  Currently XGL200 holds onto the speed value, while MGlide holds onto the compatibility factor.

Comparrision of XGL200 v.04 and MGlide v1.246G: This section is a comparison on the two wrappers and how well the performed in certain situations.  All these test were run on a PII300, 64MB RAM, Diamond Viper V550 w/1.88 Detonator Drivers, and Soundblaster PCI64.  UltraHLE is also an extremely CPU intensive program, meaning the limiting factor in the benchmarks is the CPU and not the videocard.  The videocard does also play a vital role in the benchmark, but not nearly as much as the CPU.


XGL200 v.04 Tests
Settings - TNT Default Settings w/ Real-Time Priority
UltraHLE patched with UltraHLP using Ultra ini version 1.0.16
All UltraHLE Tests with Sound Enabled
Test Run Benchmark Result Pictures
Mario64 @ 1024x768 26.6FPS
Zelda64 @ 1024x768 7.7FPS
Starfox64 @ 1024x768 14.8FPS
NBA Live 98 N/A
NHL 98 N/A FAILED TO RUN
Uprising N/A FAILED TO RUN

MGlide v1.246G Tests
Settings - W Buffer, Default TMU Memory, Middle Palette Emulation
UltraHLE patched with UltraHLP using Ultra ini version 1.0.16
All UltraHLE Tests with Sound Enabled
Test Run Benchmark Result Pictures
Mario64 @ 1024x768 15.2FPS
Zelda64 @ 1024x768 7.1FPS
Starfox64 @ 1024x768 10.2FPS
NBA Live 98 N/A
NHL 98 N/A FAILED TO RUN
Uprising N/A FAILED TO RUN

As you can tell by the benchmarks from UltraHLE, XGL200 v.04 is faster than MGlide v1.246G.  There is one main reason, because of the priority levels.  The priority allows the CPU to be more controlling towards XGL200 giving it a certain edge.  Both wrappers have very good image quality, but MGlide edges out XGL200 by a tiny little bit.  You can tell by the better blended objects and slightly better shadowing.  But in NBA Live 98, both wrappers quite frankly, aren't too hot.  Since NBA Live 98 uses Paletted Textures, and the TNT doesn't support paletted textures, the graphics will not appear correctly.  But, both wrappers have paletted texture emulation, but both seem to use the same exact form of doing this texture format because you can tell from the screenshot that they share the same exact problem.  None of the players have textures and a lot of the court is messed up.  They all appear a nasty looking orange color.

Note in the UltraHLE benchmarks, UltraHLE uses the CPU for all sound mixing and output, putting even more strain on the CPU.  Because of the lower frame rates, the sound skipping on MGlide is much worse.  Mario64 on XGL200 is perfect, crystal clear and never crackles or skips, but on MGlide, sound skipping is persistant, but not entirely bothersome.  On Zelda64, skipping is extremely frequent on both wrappers, it would be best to turn the sound OFF on this game for either wrapper.  On Starfox, it is quite an odd story, the sound skipping depends on the scene of the game.  On XGL200 sound skipping isn't too bad during speaches and talking, but in game it tends to skip a little bit (around every 5 seconds), on MGlide, skipping is very consistant to the point where the sound is very annoying.

The speed on the wrappers on UltraHLE is also a major factor when considering benchmarks.  Mario64 is very fast XGL200 and runs very smooth.  On MGlide however, the speed is a tad slower, and not as smooth, but still maintains a playable frame rate that is tolerable.  On Zelda64, the game speed is very pathetic, mainly due to the fact that the game is so large (256mbit cart, 32MB) and sucks up all the RAM and CPU power.  Frame rates and playable speeds are extremely slow and difficult to play in.  On Starfox64, the game has decent speed, but not upto a full N64 speed on either wrapper.  Speed is more tolerable on XGL200, but the game is extremely slow in the map screens on XGL200, while map screens on MGlide are a lot more smooth.

Compatibility is another factor when considering what wrapper to use.  Unfourunately, neither wrapper could run Uprising or NHL 98 at all.  In the readme file of MGlide it appears that MGlide has better compatibility than XGL200 does.  Once these other games start working on the wrappers, that'll be great.  Till then, we wait for another great wrapper to come out.

Conclusion: To conclude, currently XGL200 would be my wrapper of choice because of it's higher speed, but MGlide has slightly better image quality and compabitility.  Use which ever wrapper best suits your computer.  If you have a very fast computer (i.e. PIII450, 128mb ram, TNT2 card), then use MGlide, as it'll provide better compatibility and has SIMD KNI optimizations which will provide a better speed increase.  MGlide also has good 3DNow! optimizations so a AMD 3DNow! enhanced processor will be better with MGlide.  But if you use a slower machine (PII266, 64mb ram, TNT card), then stick with XGL200 because it is a bit faster than MGlide.

Any mistakes in this document please send corrections to Chris Kim so I can fix these mistakes ASAP, reports are very appreciated.


Added:  Wednesday, June 16, 1999
Reviewer:  Chris Kim
Score:

 1  

[ Back to reviews index ]

Home :: Share Your Story
Site contents copyright Glide Underground.
Want to syndicate our news? Hook in to our RSS Feed.