HOWTO: Buy a video card for GTA IV/EFLC

[crackdawg]

This is a simple formula for buying a video card that has predictable performance while saving time and money.

 

To prove it works I'm only going to use the information and resources that were presented when trying to debunk it except at top where I just give links for clarification.

 

I don't give links to threads because they are almost 99% people literally just saying I'm wrong for social reasons, or just simply saying "You're wrong" in entire post(s).

 

First off..why it works without my own explanation:

http://www.gpureview.com/texture-fill-rate-article-375.html

http://en.wikipedia.org/wiki/Deferred_shading

http://forum.beyond3d.com/showthread.php?t=47508

http://www.overclock.net/faqs/28489-info-w...-pipelines.html

http://www.videocardbenchmark.net/gpu_list.php

http://en.wikipedia.org/wiki/Texture_mapping_unit

 

The purchasing formula for decent FPS at XGA resolution or better with medium to very high settings:

768+ VRAM and 30+ TFR

 

The well established calculation for getting TFR:

Texture Fill Rate = (# of TMUs(texture mapping units)) x (Core Clock)

 

Here is an entire post, and one link given by my critics that inadvertently show it actually works. If I missed any links please post them here.

 

The only note I have is that this first submission reflects it all the way through except the last one where you see the core and memory clocks were changed, but the TFR stays the same. If anyone knows how you under-clock core clock and keep the same TFR please post the explanation here...TMU cant be changed... I asked it on ocerclockers IRC and even they said thats not possible..trying to alter the results?

 

 

After reading this thread, which a debate has started.

One member states GTA IV's primary performance advantage is based on TFR and VRAM while some say otherwise.

Let's set this straight and find out.

 

Note: For those that don't know what is TFR and VRAM

TFR: Texture Fill Rate

VRAM: Video Memory

e.g. If you got a video card which has 512MB, the amount of VRAM you have is 512MB

 

 

Test bed (It's in my sig)

Intel Xeon X3350 @ 2.66GHz (stock clocks)

Asus ROG board based on the Intel X48 chipset

4GB DDR2 800MHz (Kingston Hyper X)

Western Digital Black 500GB SATA 3Gb/s

ATi Radeon HD5850 (Catalyst 10.9) on PCIe x16 @full x16

Creative X-FI Xtreme Music

Vista Ultimate x64 w/SP2

24" @ 1920x1200

Game: Grand Theft Auto IV patched 1.0.7.0

 

All test ran at these settings:

 

Test #1

Using ATi's OverDrive tuning utility I've set both the GPU and RAM clocks to its lowest levels.

This brings the TFR down to 39.6

 

 

Test #1 results

 

StatisticsAverage FPS: 34.61Duration: 37.53 secCPU Usage: 79%System memory usage: 89%Video memory usage: 81%Graphics SettingsVideo Mode: 1920 x 1200 (60 Hz)Texture Quality: HighShadow Quality: HighReflection Resolution: HighWater Quality: HighTexture Filter Quality: Anisotropic x16Night Shadows: HighView Distance: 50Detail Distance: 50HardwareMicrosoft® Windows Vista" Ultimate Service Pack 2Video Adapter: ATI Radeon HD 5800 Series Video Driver version: 8.17.10.1043Audio Adapter: Speakers (Creative SB X-Fi)Intel® Xeon® CPU           X3350  @ 2.66GHzFile ID: Benchmark.cli

 

 

Test #2

This time, the clocks are set back to its factory defaults.

TFR is now at 52.2

 

 

Test #2 results

 

StatisticsAverage FPS: 45.40Duration: 37.38 secCPU Usage: 88%System memory usage: 81%Video memory usage: 81%Graphics SettingsVideo Mode: 1920 x 1200 (60 Hz)Texture Quality: HighShadow Quality: HighReflection Resolution: HighWater Quality: HighTexture Filter Quality: Anisotropic x16Night Shadows: HighView Distance: 50Detail Distance: 50HardwareMicrosoft® Windows Vista" Ultimate Service Pack 2Video Adapter: ATI Radeon HD 5800 Series Video Driver version: 8.17.10.1043Audio Adapter: Speakers (Creative SB X-Fi)Intel® Xeon® CPU           X3350  @ 2.66GHzFile ID: Benchmark.cli

 

 

Test #3

Set the graphic card to its max.

TFR at 55.8

 

 

Test #3 results

 

StatisticsAverage FPS: 45.46Duration: 37.37 secCPU Usage: 87%System memory usage: 81%Video memory usage: 81%Graphics SettingsVideo Mode: 1920 x 1200 (60 Hz)Texture Quality: HighShadow Quality: HighReflection Resolution: HighWater Quality: HighTexture Filter Quality: Anisotropic x16Night Shadows: HighView Distance: 50Detail Distance: 50HardwareMicrosoft® Windows Vista" Ultimate Service Pack 2Video Adapter: ATI Radeon HD 5800 Series Video Driver version: 8.17.10.1043Audio Adapter: Speakers (Creative SB X-Fi)Intel® Xeon® CPU           X3350  @ 2.66GHzFile ID: Benchmark.cli

 

 

Test #4

Decided to do some further research, this time the GPU to its max while RAM to its lowest.

TFR at 55.8

 

 

Test #4 results

 

StatisticsAverage FPS: 39.51Duration: 37.46 secCPU Usage: 85%System memory usage: 79%Video memory usage: 81%Graphics SettingsVideo Mode: 1920 x 1200 (60 Hz)Texture Quality: HighShadow Quality: HighReflection Resolution: HighWater Quality: HighTexture Filter Quality: Anisotropic x16Night Shadows: HighView Distance: 50Detail Distance: 50HardwareMicrosoft® Windows Vista" Ultimate Service Pack 2Video Adapter: ATI Radeon HD 5800 Series Video Driver version: 8.17.10.1043Audio Adapter: Speakers (Creative SB X-Fi)Intel® Xeon® CPU           X3350  @ 2.66GHzFile ID: Benchmark.cli

 

 

Test #5

The opposite is set: RAM and its highest, while GPU at its lowest

TFR: 39.6

 

 

Test #5 results

 

StatisticsAverage FPS: 41.82Duration: 37.45 secCPU Usage: 87%System memory usage: 79%Video memory usage: 81%Graphics SettingsVideo Mode: 1920 x 1200 (60 Hz)Texture Quality: HighShadow Quality: HighReflection Resolution: HighWater Quality: HighTexture Filter Quality: Anisotropic x16Night Shadows: HighView Distance: 50Detail Distance: 50HardwareMicrosoft® Windows Vista" Ultimate Service Pack 2Video Adapter: ATI Radeon HD 5800 Series Video Driver version: 8.17.10.1043Audio Adapter: Speakers (Creative SB X-Fi)Intel® Xeon® CPU           X3350  @ 2.66GHzFile ID: Benchmark.cli

 

 

 

Summary/Conclusion

 

1. TFR is affected by the clock speed of the graphic's GPU.

 

2. TFR does not have any advantage towards GTA IV. In fact it's the opposite.

We can see this clearly in Test #4.

 

3. GTA IV is affected by the speed of your graphic memory as stated in Test #5.

 

Note: Amount of VRAM does not equal to clock speed

 

 

The signicant numbers from test:

 

TFR-FPS 1680X105039.6:34.6152.2:45.4055.8:45.4655.8:39.51(775/1000 to 550/1125)

 

 

UPDATE: Yet another source(cross-reference with wikipedia tables below)

 

Another link that was submitted that also shows it works..no note needed:

http://www.pcgameshardware.com/aid,743498/...eviews/?page=11

 

NOTE: If you're going to critisize it please only post technical reasons or links. Please no social slander or trolling in this one.

 

EDIT: This should help a lot of people:

http://en.wikipedia.org/wiki/Comparison_of...rocessing_units

http://en.wikipedia.org/wiki/Comparison_of...rocessing_units

Edited October 30, 2010 by crackdawg

[mkey82]

Also people, note that if you find two cards with same TFR the card with higher VRAM access speed will give you better performance.

[crackdawg]

 

Also people, note that if you find two cards with same TFR the card with higher VRAM access speed will give you better performance.

TFR is TMU count multiplied by core speed, feel free to show us how VRAM clock alters TMU configuration at any time..

 

I actually noted that when I made the post. Everything except that shows the formula is true.

 

BTW x360 is 30FPS and PS3 is 22FPS at 1080p in the heaviest loads in GTA IV.

 

Next try please.

 

EDIT:"higher access speed"...bwahahaha what do you think TMU is? Please learn how they work before trying to debunk, but thanks for yet again confirming my formula.

 

EDIT: Why did you delete the post if you know so much?

Edited October 30, 2010 by crackdawg

[mkey82]

EDIT: Why did you delete the post if you know so much?

What did I delete? And how does one delete posts on these boards? Maybe I'm in cahoots with a mod or an admin lol

 

Seriously, you're getting paranoid.

[crackdawg]

EDIT: Why did you delete the post if you know so much?

What did I delete? And how does one delete posts on these boards? Maybe I'm in cahoots with a mod or an admin lol

 

Seriously, you're getting paranoid.

It's possible based on time and if there are other responses, but I actually was looking at something else and should of paraphrased yours.

 

No need to be paranoid when you can just sit back and have your critics bring you more content that reflects what you're claiming directly. You even said increase VRAM clock which compensates on cards with less TMUs(1/2 what calculates TFR).

 

At the end of the day I don't even really have to defend it, your groups 'proof' all contributes to my claim, and outside 4(?) actual technical posts the rest of your sides input on the subject is literally just slander, like bringing my intelligence into question when you have nothing to discredit it except social level attacks through irrelevant posts all made by the same ~5 people..

[Garfield 2]

Also people, note that if you find two cards with same TFR the card with higher VRAM access speed will give you better performance.

TFR is TMU count multiplied by core speed, feel free to show us how VRAM clock alters TMU configuration at any time..

 

I actually noted that when I made the post. Everything except that shows the formula is true.

 

BTW x360 is 30FPS and PS3 is 22FPS at 1080p in the heaviest loads in GTA IV.

 

Next try please.

 

EDIT:"higher access speed"...bwahahaha what do you think TMU is? Please learn how they work before trying to debunk, but thanks for yet again confirming my formula.

 

EDIT: Why did you delete the post if you know so much?

GTA IV is less than 22fps at heavy loads as I can get it down to unplayable. If Rockstar weren't money whores, they would've fixed the Xbox 360 version of GTA IV for the PS3...

[crackdawg]

Also people, note that if you find two cards with same TFR the card with higher VRAM access speed will give you better performance.

TFR is TMU count multiplied by core speed, feel free to show us how VRAM clock alters TMU configuration at any time..

 

I actually noted that when I made the post. Everything except that shows the formula is true.

 

BTW x360 is 30FPS and PS3 is 22FPS at 1080p in the heaviest loads in GTA IV.

 

Next try please.

 

EDIT:"higher access speed"...bwahahaha what do you think TMU is? Please learn how they work before trying to debunk, but thanks for yet again confirming my formula.

 

EDIT: Why did you delete the post if you know so much?

GTA IV is less than 22fps at heavy loads as I can get it down to unplayable. If Rockstar weren't money whores, they would've fixed the Xbox 360 version of GTA IV for the PS3...

I don't doubt it.

 

Most games don't have near the computation as IV, you can even see it from looking at the ped scripts. I'd also like to see IV run under a tool that analyzes memory maps for memory leaks.

[mkey82]

From what I observed on my friends PS3, the resolution is nowhere NEAR 1080p (do you even consider what you're writing sometimes?), well maybe when looking from three meter distance, maybe you couldn't see the lack of detail. Also, the FPS definitively gets into single digits area when a chain reaction explosion sets in. 22FPS stable would be a good thing for the PS3.

[crackdawg]

The blur effects and strange looking trees are actually the effect of R* custom shaders. They even addressed it in an interview once.

 

I use to exclusively play on PS3 for the first 6 months the game was released. I never experienced anything like unresponsive input or pop-ins while racing at high speeds, maybe occasional pop-ins that filled before I passed. It did okay though, and I didn't get killed in any PVP matches even in helicopters in city with people shooting rockets.

 

I'm mainly just saying you can use TFR to buy a video card for it on PC and get decent performance as a result. Nothing presented has actually proven that theory wrong, actually just the opposite. In the entire spectrum of video card specs core clock, TMU count, and shader clock(which also uses TMU count for mapping in VRAM) is what matters as even your sides 'proof'(which I actually used to publicly demonstrate) shows.

 

"(do you even consider what you're writing sometimes?)" Obviously I do, your side has failed on theoretical, applied, and technical levels to present anything so far that I couldn't use here to verify it works. Also that's called slander and it was actually in your sides posts in response way before I even started..

Edited October 31, 2010 by crackdawg

[marmoo]

well I'm glad that is cleared up

[mkey82]

"(do you even consider what you're writing sometimes?)" Obviously I do, your side has failed on theoretical, applied, and technical levels to present anything so far that I couldn't use here to verify it works. Also that's called slander and it was actually in your sides posts in response way before I even started..

Well, if you think PS3 manages to render GTAIV in 1080p you need some sort of professional advice.

[viometrix]

crackdawg,

 

again you are an idiot! you asked mkey what good is faster vram gonna do? what the f*ck.... nah dude lets just throw some old 133 mhz sd ram on there itll do just fine. faster ram means faster texture fill rates and faster per pixel shader changes. you really dont know what you are talking about. you use other peoples graphs and hardwork as your own, yet you complain at everyone that doesnt provide a graph is wrong. if we lived anywhere close id drive over and beat your fat pizza sucking ass. yes you are just that annoying. if you worked for my comapny id fire you so quick. if you owned a company like mine you wouldnt have had 25 successful years, you had a failed company in 5 years due to your honest lack of any true hardcore computer knowledge. ive had retarded kids work for me through an education program that are smarter then you!

[crackdawg]

 

crackdawg,

 

again you are an idiot! you asked mkey what good is faster vram gonna do? what the f*ck.... nah dude lets just throw some old 133 mhz sd ram on there itll do just fine. faster ram means faster texture fill rates and faster per pixel shader changes. you really dont know what you are talking about. you use other peoples graphs and hardwork as your own, yet you complain at everyone that doesnt provide a graph is wrong. if we lived anywhere close id drive over and beat your fat pizza sucking ass. yes you are just that annoying. if you worked for my comapny id fire you so quick. if you owned a company like mine you wouldnt have had  25 successful years, you had a failed company in 5 years due to your honest lack of any true hardcore computer knowledge. ive had retarded kids work for me through an education program that are smarter then you!

You mean where I clearly state it only compensates for lack of TMUs? I looked through my posts, feel free to quote me saying that at any time, even in a remote context.

 

Also read in OP where I say no trolling or slander in this one *please*..

 

You might want to learn how video cards work, and how to debate, cause all I'm seeing is a troll with no working knowledge of the subject matter.. you've literally presented nothing.

[viometrix]

no ive only owned a computer company for the last 25 years, i got here knowing nothing. and anything you say invites this sh*t, everyone has been against you, but i guess we are all wrong and you lord and master of all you see are right. you call us trolls because we dont like you. nobody here does. then again no one liked me till i learned to respect others and keep my mouth shut, so there is hope for you. but anyone here knows even without this severe need for proof you seem to need, i am ussually right 99% when it comes to computer and hardware - i never claim to be an expert in the mechanics of this game. we dont like you, so go away.

[crackdawg]

 

no ive only owned a computer company for the last 25 years, i got here knowing nothing. and anything you say invites this sh*t, everyone has been against you, but i guess we are all wrong and you lord and master of all you see are right. you call us trolls because we dont like you. nobody here does. then again no one liked me till i learned to respect others and keep my mouth shut, so there is hope for you. but anyone here knows even without this severe need for proof you seem to need, i am ussually right 99% when it comes to computer and hardware - i never claim to be an expert in the mechanics of this game. we dont like you, so go away.

I know a CEO of security firm who doesn't know how buffer overflows and XSS work, and have met an executive of an embedded IP company who didn't know how TLB worked. Saying you've owned something for 25 years means little.

 

The same 5 people following me from thread to thread presenting nothing of technical relevance are neither 'proof' nor "everyone". The ~3 people who actually did, which don't seem to be bashing but rather researching, actually presented information that reflects my formula, thus why I use it here.

 

Maybe you should spend more time running the company which is probably not known outside of a few state counties in an american state, and spend less time trolling forums comparing stuff that actually present *anything* technical, to your social status and why because you and your friends don't like me it means I'm wrong.

 

Until you can actually present even a sample of technical document or *anthing* that even suggests I'm the slightest bit wrong, how about spouting your pride elsewhere.. the same goes for your buddies who also have literally presented nothing to the argument but social slander and telling us what they've done in life..

Edited November 1, 2010 by crackdawg

[viometrix]

no ive only owned a computer company for the last 25 years, i got here knowing nothing. and anything you say invites this sh*t, everyone has been against you, but i guess we are all wrong and you lord and master of all you see are right. you call us trolls because we dont like you. nobody here does. then again no one liked me till i learned to respect others and keep my mouth shut, so there is hope for you. but anyone here knows even without this severe need for proof you seem to need, i am ussually right 99% when it comes to computer and hardware - i never claim to be an expert in the mechanics of this game. we dont like you, so go away.

I know a CEO of security firm who doesn't know how buffer overflows and XSS work, and have met an executive of an embedded IP company who didn't know how TLB worked. Saying you've owned something for 25 years means little.

 

The same 5 people following me from thread to thread presenting nothing of technical relevance are neither 'proof' nor "everyone". The ~3 people who actually did, which don't seem to be bashing but rather researching, actually presented information that reflects my formula, thus why I use it here.

 

Maybe you should spend more time running the company which is probably not known outside of a few state counties in an american state, and spend less time trolling forums comparing stuff that actually present *anything* technical, to your social status and why because you and your friends don't like me it means I'm wrong.

 

Until you can actually present even a sample of technical document or *anthing* that even suggests I'm the slightest bit wrong, how about spouting your pride elsewhere.. the same goes for your buddies who also have literally presented nothing to the argument but social slander and telling us what they've done in life..

if i dont know something i ask my techs who do. also my company is H&J international (www.askchip.com) so i am known all over the world and usa. so next time present your own data and graphs, and then the others to prove things. without your own data and hard work to show (not just someone elses) why would anyone belive you?

[crackdawg]

 

no ive only owned a computer company for the last 25 years, i got here knowing nothing. and anything you say invites this sh*t, everyone has been against you, but i guess we are all wrong and you lord and master of all you see are right. you call us trolls because we dont like you. nobody here does. then again no one liked me till i learned to respect others and keep my mouth shut, so there is hope for you. but anyone here knows even without this severe need for proof you seem to need, i am ussually right 99% when it comes to computer and hardware - i never claim to be an expert in the mechanics of this game. we dont like you, so go away.

I know a CEO of security firm who doesn't know how buffer overflows and XSS work, and have met an executive of an embedded IP company who didn't know how TLB worked. Saying you've owned something for 25 years means little.

 

The same 5 people following me from thread to thread presenting nothing of technical relevance are neither 'proof' nor "everyone". The ~3 people who actually did, which don't seem to be bashing but rather researching, actually presented information that reflects my formula, thus why I use it here.

 

Maybe you should spend more time running the company which is probably not known outside of a few state counties in an american state, and spend less time trolling forums comparing stuff that actually present *anything* technical, to your social status and why because you and your friends don't like me it means I'm wrong.

 

Until you can actually present even a sample of technical document or *anthing* that even suggests I'm the slightest bit wrong, how about spouting your pride elsewhere.. the same goes for your buddies who also have literally presented nothing to the argument but social slander and telling us what they've done in life..

if i dont know something i ask my techs who do. also my company is H&J international (www.askchip.com) so i am known all over the world and usa. so next time present your own data and graphs, and then the others to prove things. without your own data and hard work to show (not just someone elses) why would anyone belive you?

Actually using your critics data(which all of that wasn't their own...) to prove your theory is suppose to prove it's validity even that much more. Like in my previous posts I'm just assuming you and your friends comprehension is so lacking that you just don't 'get it'..but every benchmark out their still reflects my data, you have literally none though so you are neither right or wrong except in your accusations which are flawed in more ways than one.

 

You and your friends may want to take your own advice when telling me to present my own data. I'm not referring to the few people who presented data when I mention people who keep posting but present nothing of technical relevance. Nobody here cares what you own or what your were doing x-years ago, it means 0 in this debate.

 

Also enlighten the community on how owning a computer shipping company that has one reference in google, in any way gives you credibility? Just a quick browse across the domain it doesn't look like it's ever done well, and the site and stock you have alone is pretty outdated. You ship stuff...

 

You continue to make posts that present literally nothing but social attacks..sorry but your haggard shipping company means nothing here hot shot.

 

EDIT: BTW..in the small NC town in America where my parents live, a former gym coach started a company exactly like yours two years ago, and I'm pretty sure he rakes in more business already. Like I said...irrelevant.

Edited November 2, 2010 by crackdawg

[viometrix]

lmfao rofl

[crackdawg]

lmfao rofl

To put it in better perspective, that post presents about the same amount of technical data as your others..literally. Tha'ts minimal enough for your lacking comprehensions, hopefully.

 

post something relevant or go troll another thread..

[viometrix]

you are quite talked about around here in your own regards - your entertainment value is limitless.... lmao rofl, but ive had enough of this childish debate, good day

Edited November 3, 2010 by viometrix

[crackdawg]

 

you are quite talked about around here in your own regards - your entertainment value is limitless.... lmao rofl, but ive had enough of this childish debate, good day

I'm sure..at the end of the day though I won the technical aspect of the debate, even using the critics data to show my formula. What's more is even using my formula to actually buy a card shows it works.

 

The only ~4 people who presented information left the debate a while ago. You and the other trolls are all that is left, and anyone who considers your slander to have some merit should go look through you, and your friends posts, and see what kind of department store sale department garbage you spout here to people looking for hardware.

 

In this argument I presented data, you told us about your company which is what appears to be a warehouse of 3+ year old computer surplus which somehow gives you credibility in the field of computer engineering..

 

EDIT: btw this isn't the first time I've attracted haters here. I've argued with people here who actually had talent, mostly in RCE and programming, and proved my side clearly there too. You're obviously mad at me cause I came up with this formula which makes most of your posts look both irrelevant and noob. Saying people are talking about me is irrelevant, they always have because I'm educated and they/you are bitter burnouts.

 

It also pays to point out that I'm a software/computer engineer for a living for a notable company, and you run a computer surplus warehouse full of dated hardware, that has lot prices and surplus that only morons would invest in. I can use the social classification in my favor too. I actually got somewhere in the industry as a result of talent.

Edited November 3, 2010 by crackdawg

[viometrix]

you are quite talked about around here in your own regards - your entertainment value is limitless.... lmao rofl, but ive had enough of this childish debate, good day

I'm sure..at the end of the day though I won the technical aspect of the debate, even using the critics data to show my formula. What's more is even using my formula to actually buy a card shows it works.

 

The only ~4 people who presented information left the debate a while ago. You and the other trolls are all that is left, and anyone who considers your slander to have some merit should go look through you, and your friends posts, and see what kind of department store sale department garbage you spout here to people looking for hardware.

 

In this argument I presented data, you told us about your company which is what appears to be a warehouse of 3+ year old computer surplus which somehow gives you credibility in the field of computer engineering..

 

EDIT: btw this isn't the first time I've attracted haters here. I've argued with people here who actually had talent, mostly in RCE and programming, and proved my side clearly there too. You're obviously mad at me cause I came up with this formula which makes most of your posts look both irrelevant and noob. Saying people are talking about me is irrelevant, they always have because I'm educated and they/you are bitter burnouts.

 

It also pays to point out that I'm a software/computer engineer for a living for a notable company, and you run a computer surplus warehouse full of dated hardware, that has lot prices and surplus that only morons would invest in. I can use the social classification in my favor too. I actually got somewhere in the industry as a result of talent.

and i make millions of dollars a year adjusting to the economy. i sell what people can afford, the rest is servers and custom builds. and you you know whether you are actually right or wrong at the end of the day doesnt matter when you are so stuck up your head loops around and ends up up your ass. also i remember seeing somewhere back you said you where a different engineer not directly in the computer field, ill have to find that post.

 

but you know something, i can rest assured no matter how right you may think you are, at the end of the day it all comes down to this - (heres me being an asshole snob like you) with my 2bit distribution wharehouse as you like to call it, i personally make more in one month then you make in a whole year. i dont have to work for a reputable company under a boss, i am the boss. so big guy, he who dies with the most toys wins. that will be me :-)

 

now stop responding and having to have the last word, makes you look childish.

[TFRFTWTFRFTW]

Im an idiot.

[okei]

What the f*ck is going on?

[TFRFTWTFRFTW]

What the f*ck is going on?

LOL LOL LOL

[viometrix]

What the f*ck is going on?

personally i think it is crackdawg being he was so passionate about the very subjuct. let the mods sort it out and see if it came from the same ip or one remotely close. hopefully getting him banned for stupidity and multiple accounts.

[TFRFTWTFRFTW]

What the f*ck is going on?

personally i think it is crackdawg being he was so passionate about the very subjuct. let the mods sort it out and see if it came from the same ip or one remotely close. hopefully getting him banned for stupidity and multiple accounts.

I'm not Crackdawg LOL, just having a lil fun. If you find it annoying then I'll stop. Peace man

[viometrix]

What the f*ck is going on?

personally i think it is crackdawg being he was so passionate about the very subjuct. let the mods sort it out and see if it came from the same ip or one remotely close. hopefully getting him banned for stupidity and multiple accounts.

I'm not Crackdawg LOL, just having a lil fun. If you find it annoying then I'll stop. Peace man

thank you

[Anasterian]

This thread...

 

 

... is fun.

 

 

[Stinky12]

After watching this thread for some time, it concludes, OP does not know what he's talking about.

Here are my evidence for that statement.

 

Evidence #1

 

 

Because RAGE engine has it's own rendering pipeline, and only uses texture and framebuffer features of your card and subsystem. This is why I say buy by TFR and VRAM, not what is new or model numbers, or even clocks like all the experts here like to repeatedly suggest.

 

source in 5th post

 

In the quoted statement above, it says "buy according to TFR and VRAM, not what is new or model numbers or even clocks".

 

Then we have this quote:

 

The well established calculation for getting TFR:

Texture Fill Rate = (# of TMUs(texture mapping units)) x (Core Clock)

 

source in 1st post

 

If TFR and VRAM is all that matters, not based on "clocks", then why does OP then say TFR = TMU x Core "Clocks"?

Does it matter what your GPU (Core Clock) is running at if the first quote statement is true?

 

Evidence #2

 

 

http://www.pcgameshardware.com/aid,743498/...eviews/?page=11

So, how will you explain that? Here we can see that GTX470 (which, mind you, has actually TFR = 34 - even lower than GTX460) is more than twice as fast as 8800GTX (which has TFR = 36.8 - higher than 8800GT).

At 1680x1050 8800 gets 30.6 FPS avg and 480 72.7. 480 is also 4.2 TFR over the 460. I'm not sure how my posts are inconsistent? 8800GT and the 460 aren't even there, but the data still validates my statements.

 

This post has been edited by crackdawg on Oct 25 2010, 17:16

8800GTX has higher TFR than GTX470, but performs noticeably worse - that's what I said. What's in your opinion the reason behind this? Something that also discredits that sites benchmark:

470 GTX "March 26, 2010"

8800 GTX "November 8, 2006"

"Geforce 197.13/17(Q)"

 

source

 

 

In contradiction, we have

 

Another link that was submitted that also shows it works..no note needed:

http://www.pcgameshardware.com/aid,743498/...eviews/?page=11

 

source

 

 

If OP discredits the link yojo2 provided, then why does he use the same exact link to prove his point?

This does not make any sense!

 

 

Evidence #3

 

 

The only note I have is that this first submission reflects it all the way through except the last one where you see the core and memory clocks were changed, but the TFR stays the same. If anyone knows how you under-clock core clock and keep the same TFR please post the explanation here...TMU cant be changed... I asked it on ocerclockers IRC and even they said thats not possible..trying to alter the results?

 

 

After reading this thread, which a debate has started.

One member states GTA IV's primary performance advantage is based on TFR and VRAM while some say otherwise.

Let's set this straight and find out.

 

Note: For those that don't know what is TFR and VRAM

TFR: Texture Fill Rate

VRAM: Video Memory

e.g. If you got a video card which has 512MB, the amount of VRAM you have is 512MB

 

 

Test bed (It's in my sig)

Intel Xeon X3350 @ 2.66GHz (stock clocks)

Asus ROG board based on the Intel X48 chipset

4GB DDR2 800MHz (Kingston Hyper X)

Western Digital Black 500GB SATA 3Gb/s

ATi Radeon HD5850 (Catalyst 10.9) on PCIe x16 @full x16

Creative X-FI Xtreme Music

Vista Ultimate x64 w/SP2

24" @ 1920x1200

Game: Grand Theft Auto IV patched 1.0.7.0

 

All test ran at these settings:

 

Test #1

Using ATi's OverDrive tuning utility I've set both the GPU and RAM clocks to its lowest levels.

This brings the TFR down to 39.6

 

 

Test #1 results

 

StatisticsAverage FPS: 34.61Duration: 37.53 secCPU Usage: 79%System memory usage: 89%Video memory usage: 81%Graphics SettingsVideo Mode: 1920 x 1200 (60 Hz)Texture Quality: HighShadow Quality: HighReflection Resolution: HighWater Quality: HighTexture Filter Quality: Anisotropic x16Night Shadows: HighView Distance: 50Detail Distance: 50HardwareMicrosoft® Windows Vista" Ultimate Service Pack 2Video Adapter: ATI Radeon HD 5800 Series Video Driver version: 8.17.10.1043Audio Adapter: Speakers (Creative SB X-Fi)Intel® Xeon® CPU           X3350  @ 2.66GHzFile ID: Benchmark.cli

 

 

Test #2

This time, the clocks are set back to its factory defaults.

TFR is now at 52.2

 

 

Test #2 results

 

StatisticsAverage FPS: 45.40Duration: 37.38 secCPU Usage: 88%System memory usage: 81%Video memory usage: 81%Graphics SettingsVideo Mode: 1920 x 1200 (60 Hz)Texture Quality: HighShadow Quality: HighReflection Resolution: HighWater Quality: HighTexture Filter Quality: Anisotropic x16Night Shadows: HighView Distance: 50Detail Distance: 50HardwareMicrosoft® Windows Vista" Ultimate Service Pack 2Video Adapter: ATI Radeon HD 5800 Series Video Driver version: 8.17.10.1043Audio Adapter: Speakers (Creative SB X-Fi)Intel® Xeon® CPU           X3350  @ 2.66GHzFile ID: Benchmark.cli

 

 

Test #3

Set the graphic card to its max.

TFR at 55.8

 

 

Test #3 results

 

StatisticsAverage FPS: 45.46Duration: 37.37 secCPU Usage: 87%System memory usage: 81%Video memory usage: 81%Graphics SettingsVideo Mode: 1920 x 1200 (60 Hz)Texture Quality: HighShadow Quality: HighReflection Resolution: HighWater Quality: HighTexture Filter Quality: Anisotropic x16Night Shadows: HighView Distance: 50Detail Distance: 50HardwareMicrosoft® Windows Vista" Ultimate Service Pack 2Video Adapter: ATI Radeon HD 5800 Series Video Driver version: 8.17.10.1043Audio Adapter: Speakers (Creative SB X-Fi)Intel® Xeon® CPU           X3350  @ 2.66GHzFile ID: Benchmark.cli

 

 

Test #4

Decided to do some further research, this time the GPU to its max while RAM to its lowest.

TFR at 55.8

 

 

Test #4 results

 

StatisticsAverage FPS: 39.51Duration: 37.46 secCPU Usage: 85%System memory usage: 79%Video memory usage: 81%Graphics SettingsVideo Mode: 1920 x 1200 (60 Hz)Texture Quality: HighShadow Quality: HighReflection Resolution: HighWater Quality: HighTexture Filter Quality: Anisotropic x16Night Shadows: HighView Distance: 50Detail Distance: 50HardwareMicrosoft® Windows Vista" Ultimate Service Pack 2Video Adapter: ATI Radeon HD 5800 Series Video Driver version: 8.17.10.1043Audio Adapter: Speakers (Creative SB X-Fi)Intel® Xeon® CPU           X3350  @ 2.66GHzFile ID: Benchmark.cli

 

 

Test #5

The opposite is set: RAM and its highest, while GPU at its lowest

TFR: 39.6

 

 

Test #5 results

 

StatisticsAverage FPS: 41.82Duration: 37.45 secCPU Usage: 87%System memory usage: 79%Video memory usage: 81%Graphics SettingsVideo Mode: 1920 x 1200 (60 Hz)Texture Quality: HighShadow Quality: HighReflection Resolution: HighWater Quality: HighTexture Filter Quality: Anisotropic x16Night Shadows: HighView Distance: 50Detail Distance: 50HardwareMicrosoft® Windows Vista" Ultimate Service Pack 2Video Adapter: ATI Radeon HD 5800 Series Video Driver version: 8.17.10.1043Audio Adapter: Speakers (Creative SB X-Fi)Intel® Xeon® CPU           X3350  @ 2.66GHzFile ID: Benchmark.cli

 

 

 

Summary/Conclusion

 

1. TFR is affected by the clock speed of the graphic's GPU.

 

2. TFR does not have any advantage towards GTA IV. In fact it's the opposite.

We can see this clearly in Test #4.

 

3. GTA IV is affected by the speed of your graphic memory as stated in Test #5.

 

Note: Amount of VRAM does not equal to clock speed

 

 

The signicant numbers from test:

 

TFR-FPS 1680X105039.6:34.6152.2:45.4055.8:45.4655.8:39.51(775/1000 to 550/1125)

 

 

Q1: OP says core and memory clock were altered but TFR stayed the same in the last test (Test #5)

Stay the same to what other test? However in Test #5 we have the same TFR rating (39.6) as Test #1.

This is because in Test #1 and #5 the GPU are clocked at the same speed of 550MHz.

 

Q2: In the "code", it says ran at a resolution of 1680x1050.

Where in those images says those test ran at the resolution.

It clearly show the test was conducted at a much higher resolution of 1920x1200

Maybe OP just made a typo?

 

Q3: Continuing with the "code", it says the following:

 

55.8:45.4655.8:39.51(775/1000 to 550/1125)

 

 

In Test #3, with a TFR at 55.8, the core clock was set at 775MHz and ram clock was set at 1125MHz

Where does it say in Test #3, the ram clock was set at 1000MHz?

 

In Test #4, also with a TFR at 55.8, the core clock is still the same at 775MHz and the ram clock is at 500MHz.

Where in it, does it say the core clock is set at 550MHz and ram clock at 1125MHz?

Could this be another typo from OP or does it strongly state OP does not know what he's talking about and not only that, doesn't have the slightest clue as to

what he was doing?

 

 

side note

 

In Test # 3 and 4, both TFR are at 55.8 and GPUs are clock the same at 775MHz, one shows

with a FPS of 45.46 and the other 39.51. Why is there such a difference if both core clock are set at the same speed?

 

This is because of the ram clock as it affect the memory bandwidth. The bandwidth along with other important aspects to a graphic card gives out the total FPS.

RAM clocked at 500MHz and you get a bandwidth of 64GB/s

RAM clocked at 1125MHz and you get a bandwidth of 144GB/s

 

 

Lesson

Always listen to a expert as 99% of them they're always right.

When buying a graphic card for GTA IV/EFLC look at Stream Procesors, Core/RAM clock, memory bus, and other important aspects.

And TFR isn't one of them.