Nvidia Pascal GPU has 17 billion transistors

Factum said:
Like I said....there are milllions of uninformed gamers...you seem to fit the description.

If 3Dmark had any real world value...why do you think sites benchmark games..and not just put out a 3Dmark nuber and call it a day?

Here is a little history for you about 3Dmark:
http://hardforum.com/showthread.php?t=816850

3Dmark has zero corrolation with real world gaming performance, like it or not.

TL;DR
They use 3Dmark because of uninformed posters like your self.
Click to expand...


stop going back to 2004-2006 we are in 2015 and it does give the user correlation where his or her system stands as a performance indicator for gaming. Im not saying 3dmark is the only thing that matters when it comes to benchmarking a system how a game functions with a particular video card. I have mentioned before that users should look at how the video card benches on certain games.
 
Last edited:
Drakensoul said:
Just because you don't understand something doesn't make it nonsensical. :eek:

3DMark is useless to me. Telling me that a Kingpin card gets 25,000 3DMark points tells me exactly that: a Kingpin card gets 25,000 3DMark points. It means nothing. Show me a comparison between the Kingpin card and another 980Ti running Witcher3 and I can actually use that information to decide whether or not the Kingpin card is worth buying.

Very simple. :)

But again, you are entitled to your opinion. Perhaps we are all just too simpleton to understand the braintrust that you bring to the discussion. Snicker.
Click to expand...

If you cant understand 3dmark then you have issues. Because millions of people use it to give a baseline how your system performs with other users systems and video cards.
If you had any logic the kingpin 980ti would perform better than a regular non reference card due to the engineering that goes into creating the kingpin card in the first place. What EVGA is trying to demonstrate is that their card has potential to perform better which thus give you a higher frame rate .
Again i take it you have looked at comparisons between a reference gpu vs a non reference gpu? Well kingpin is some would say beyond non reference with its unique engineering that goes into the card to make it perform better than all those other non reference cards out there another example just like the MSI lightning . Also when combined with a modded bios you can believe the card will score a higher 3dmark than any card made which will ultimately give you a higher FPS in games.
Also i think i made myself clear if you cant read that everyone else should look at reviewers benchmarks in games to also see how the card operates . I can almost guarantee you that on single or SLI 3dmark scores the card with the higher 3dmark score will perform better every time.
 
Factum said:
Again, there is a lot of pagehits from millions of ignorant gamers...dosn't alter the fact that 3Ddmark is useless in comparing different SKU's.

You could use it as a primitive gauge for assessing OC improvements on a single system.

But besides that...useless.
Click to expand...

i just find this funny you are calling out millions of gamers are ignorant

Because 99% of the gaming community and professional reviewers out there would have to disagree with your statement
 
just to educate how 3dmark works

Running 3DMark produces a 3DMark score, with higher numbers indicating better performance. The 3DMark measurement unit is intended to give a normalized mean for comparing different PC hardware configurations (mostly graphics processing units and central processing units), which proponents such as gamers and overclocking enthusiasts assert is indicative of end-user performance capabilities.
Many versions of 3DMark have been released since 1998. As each 3DMark is based on a specific version of the DirectX API scores cannot be compared across the different releases.

Scores from Windows, Windows RT, Android and iOS can be compared making this version of 3DMark the first to enable cross-platform performance comparisons.
Unlike previous versions, 3DMark features separate benchmark tests with each producing its own score.
Ice Storm is a DirectX 11 feature level 9 / OpenCL ES 2.0 test targeting smartphones, tablets and entry-level PCs. Ice Storm Extreme uses more demanding settings to provide a suitable test for high-end mobile devices. Ice Storm Unlimited renders to an off-screen target to enable chip-to-chip performance comparisons free from the effects of vertical sync, display resolution scaling and other operating system factors.
Cloud Gate is a DirectX 11 feature level 10 test for typical home PC and notebooks.
Fire Strike is a DirectX 11 test for gaming PCs. Fire Strike Extreme can be used to test high-performance gaming PCs with multiple GPUs.
 
Does 3DMark tell me how much faster a GTX 980 Ti runs The Witcher 3 than a 980 does?
 
3dmark is a tool for us, using games for benchmarks are tools for us, this community is a tool for us, we are all tools...for us....;)

There, argument settled. Now back to these 17billion transistors.
 
UnrealCpu said:
I would agree

but some people on here say 3dmark is completely useless as a tool to benchmark. If anyone knows 3dmark has several tests to include physics which will max out your cards potential clock speeds. It is also a known fact the higher your 3d mark score whether it be default or extreme will net you an overall higher score than the card that has a lower gpu score. I would agree and have mentioned this before looking at reviews so see those multiple benchmarks in game is a good tool to reference how the video card performs. But to say 3dmark is useless benchmark is just dumb .
Click to expand...

Fallacy detected: Moving the goalposts.
 
evilsofa said:
Does 3DMark tell me how much faster a GTX 980 Ti runs The Witcher 3 than a 980 does?
Click to expand...

Hence the reason 3DMark is useless to me. It gives zero information useful to me. It's a flashy number which tells me nothing.

Much like putting your car on a dyno may give you flashy numbers and graphs but means nothing in terms of how the car will track.

This is a very simple concept. And someone who can't construct a paragraph telling someone they have problems and no logic and "just dumb"? :rolleyes::rolleyes::rolleyes::rolleyes::rolleyes::rolleyes:

Pleaseeeeeeeeeee boo boo. You're and your is so second grade.
 
10x.. yup and I have an inheritance from a Nigerian prince for you.
 
Drakensoul said:
Hence the reason 3DMark is useless to me. It gives zero information useful to me. It's a flashy number which tells me nothing.

Much like putting your car on a dyno may give you flashy numbers and graphs but means nothing in terms of how the car will track.

This is a very simple concept. And someone who can't construct a paragraph telling someone they have problems and no logic and "just dumb"? :rolleyes::rolleyes::rolleyes::rolleyes::rolleyes::rolleyes:

Pleaseeeeeeeeeee boo boo. You're and your is so second grade.
Click to expand...

You just want to keep dragging this on and on. Have you completed a Masters thesis yet? Because I'd like to read a 30 page paper on why 3dmark is useless to hardcorp overclockers and gamers
 
UnrealCpu said:
You just want to keep dragging this on and on. Have you completed a Masters thesis yet? Because I'd like to read a 30 page paper on why 3dmark is useless to hardcorp overclockers and gamers
Click to expand...

It's not useful because it's too easily optimized for and won't give you the correct picture on anything else.

Just like FurMark, really.

It's not even useful as a stability test anymore, lol. I find Star Swarm to be much better.
 
This is why I prefer [H] GPU reviews. I can care less about a 3dmark score when it comes to the actual games I play or will soon be playing. Real world benches is what matters to me. 3dmark doesn't tell me how my GPU will run The Witcher 3 @1440p
 
UnrealCpu said:
Running 3DMark produces a 3DMark score, with higher numbers indicating better performance. The 3DMark measurement unit is intended to give a normalized mean for comparing different PC hardware configurations (mostly graphics processing units and central processing units), which proponents such as gamers and overclocking enthusiasts assert is indicative of end-user performance capabilities.
Click to expand...

yes. so? how does this translate to the real world? two different setups can produce similar scores, but one provides a better experience to the user. what good does dick measuring do me, if all I get is a micro stutter fest?
 
UnrealCpu said:
You just want to keep dragging this on and on. Have you completed a Masters thesis yet? Because I'd like to read a 30 page paper on why 3dmark is useless to hardcorp overclockers and gamers
Click to expand...

Masters? What good is 30 pages if you can't comprehend two sentences? Oh baby... I'm so sorry :( Good luck to you in your quest for learning and intelligence. :(
 
Drakensoul said:
Masters? What good is 30 pages if you can't comprehend two sentences? Oh baby... I'm so sorry :( Good luck to you in your quest for learning and intelligence. :(
Click to expand...

Hey i can comprehend that you are a fool. So when you coming up with your hypothesis why 3dmark is usesless? I want to see at least minimum 30 pages, because i want to be the first one to read it. I will make a bet you can not even write 2 pages.:D
 
I don't need any pages, as others have posted the same thing as me already 15 times and you're just having difficulty processing it. Setting some artificial goal that if I don't write a 30 page post then nothing I say makes sense is completely asinine. Want to compare degrees? I'm perfectly happy to send you mine.

3DMark doesn't tell me anything about real-world performance. I cannot look at a 3DMark number and ascertain how a video card will perform in, say, Witcher3. Or World of Warcraft. Or FarCry, or any other game. It's just a number. It is therefore meaningless to me, and I have no use for it. If I want to know how a video card performs, I will look at a video game benchmark. From that I can glean real-world information and decide what card is worth my money. And since I can use real-world data, why would I need the 3DMark middle-man to use a synthetic number to try and model what game performance might be? I can just go straight to the source and look at the game benchmark comparisons.

If you want to spend your money based on a synthetic number, go for it. As I have already said, you are entitled to your opinion. If you want to buy your car based solely on dyno numbers, more power to you. I'll go with track times, thank you.

I'm sorry that's beyond your ability to comprehend. But if you can't manage to swallow that, what good is 30 pages going to do you? None. Best of luck. :eek:
 
Last edited:
Drakensoul said:
I don't need any pages, as others have posted the same thing as me already 15 times and you're just having difficulty processing it. Setting some artificial goal that if I don't write a 30 page post then nothing I say makes sense is completely asinine.

3DMark doesn't tell me anything about real-world performance. I cannot look at a 3DMark number and ascertain how a video card will perform in, say, Witcher3. Or World of Warcraft. Or FarCry, or any other game. It's just a number. It is therefore meaningless to me, and I have no use for it. If I want to know how a video card performs, I will look at a video game benchmark. From that I can glean real-world information and decide what card is worth my money.

I'm sorry that's beyond your ability to comprehend. Best of luck. :eek:
Click to expand...

Yeh you can... high score = better game performance 99% of the time, simple.
 
Drakensoul said:
I don't need any pages, as others have posted the same thing as me already 15 times and you're just having difficulty processing it. Setting some artificial goal that if I don't write a 30 page post then nothing I say makes sense is completely asinine.

3DMark doesn't tell me anything about real-world performance. I cannot look at a 3DMark number and ascertain how a video card will perform in, say, Witcher3. Or World of Warcraft. Or FarCry, or any other game. It's just a number. It is therefore meaningless to me, and I have no use for it. If I want to know how a video card performs, I will look at a video game benchmark. From that I can glean real-world information and decide what card is worth my money.

I'm sorry that's beyond your ability to comprehend. Best of luck. :eek:
Click to expand...

I understand what you are saying and agree . Some people say 3d mark is a useless tool which it isnt. Yeah you should use fraps etc to determine your fps in game. However using synthetic benchmarks like 3dmark will ultimately determine how your video card performs on your system and compare that to millions of other people. This information gives you a general idea how your video card is performing. I mentioned earlier combined with using fps indicators while in game is also required to determine your video cards performance. If 3dmark was useless and vally no one would be using it.
 
UnrealCpu said:
I understand what you are saying and agree . Some people say 3d mark is a useless tool which it isnt. Yeah you should use fraps etc to determine your fps in game. However using synthetic benchmarks like 3dmark will ultimately determine how your video card performs on your system and compare that to millions of other people. This information gives you a general idea how your video card is performing. I mentioned earlier combined with using fps indicators while in game is also required to determine your video cards performance. If 3dmark was useless and vally no one would be using it.
Click to expand...

People like seeing big flashy numbers! Even bigger numbers to come with Pascal's 17 billion transistors. And a FPS increase of like 5. :p
 
Nicholars said:
Yeh you can... high score = better game performance 99% of the time, simple.
Click to expand...

history shows it on every generation whether it be a new AMD or Nvidia video card . A higher 3dmark score represents a higher FPS in games and overall performance.

Its not like i dont look at in game FPS in these reviews, i do however 3dmark is not a useless tool as it can really determine the max performance of any video card versus just playing a game like witcher 3

and thats why EVGA brags about K|n|gP|n 3dmark score on every version of this particular video card when it is released to the consumer. I personally run it all the time to see my historic score if my system has improved .
 
You can guess that video card model number = higher performance 99% of the time too. That doesn't give you any relativity, and thus doesn't help you decide if an upgrade is worthwhile.

I.e.: If a 970 gets 20,000 and a 980ti gets 28,000, what does that actually mean? Nothing to me. Sounds impressive.

If the 970 gets 25fps on max settings average and the 980ti gets 30fps on max settings average, then I start to understand what the actual difference is.
 
Drakensoul said:
People like seeing big flashy numbers! Even bigger numbers to come with Pascal's 17 billion transistors. And a FPS increase of like 5. :p
Click to expand...


and when the reviews are released for the flagship pascal 3dmark scores will be 5k or more greater than an overclocked 980 ti and FPS in games like witcher 3 will increase by 20-30% or more all while using way less power. That is being conservative while pascal is slated to operate 50% or more performance vs the 980ti. I am sure this is only in certain situations while in game or at a particular resolution. I am hoping 2560x1440 will see a huge improvement since 4k is going to be the norm in a year or two.

We are going towards a new generation of video card technology . This isnt no Maxwell upgrade or a Titan cripple. We are going to see video cards with 24-36gb ram within 12 months. Maxwell is almost 2 years old now. Direct x 12 and windows 10 pascal will destroy the maxwell titan.

NVIDIA's Pascal GPU to feature over 100% more transistors than Titan X

Read more at http://www.tweaktown.com/news/46620...re-over-100-more-transistors-titan/index.html

trust me all those transistors will be put to good use and the cards will be twice as small on a video card. No more bulky video cards for a while thank god.
I cant wait until the next nvidia console is released.
 
Last edited:
Drakensoul said:
You can guess that video card model number = higher performance 99% of the time too. That doesn't give you any relativity, and thus doesn't help you decide if an upgrade is worthwhile.

I.e.: If a 970 gets 20,000 and a 980ti gets 28,000, what does that actually mean? Nothing to me. Sounds impressive.

If the 970 gets 25fps on max settings average and the 980ti gets 30fps on max settings average, then I start to understand what the actual difference is.
Click to expand...

yes i got that. I already said many times , you should look at the in game fps using fraps or whatever tool show the fps and research reviews.
Still 3dmark is not useless and should be used as a performance indicator tool
 
Drakensoul said:
So why wouldn't I just look at the game performance? You know, real world... :confused:
Click to expand...

Because you use 3Dmark to overclock and test your system is working properly... that is the point of it... you check your system is working properly and your OC is stable with 3Dmark and other benchmarking apps... that is the point of them. Its a very useful app to check your system is working as it should, test overclocks and stability etc. It is ridiculous to say that 3Dmark is useless / pointless, you run the benchmark, it gives you a score, which you can compare to similar PC's... Why is that useless?!
 
Last edited:
If I was the type to make up and spread rumors I'd probably say that Parker is going to be dropping real soon.

But I'm not the type, so I won't bring it up.
 
Parker is going to be dropping real soon. You heard it here first folks.

:D

Also, I nearly ran out of popcorn reading the last 100 or so posts.
 
Nicholars said:
Because you use 3Dmark to overclock and test your system is working properly... that is the point of it... you check your system is working properly and your OC is stable with 3Dmark and other benchmarking apps... that is the point of them. Its a very useful app to check your system is working as it should, test overclocks and stability etc. It is ridiculous to say that 3Dmark is useless / pointless, you run the benchmark, it gives you a score, which you can compare to similar PC's... Why is that useless?!
Click to expand...

Well since I don't overclock that's gone. And the rest is pointless because it gives me no real world information. I don't really care about how many 3DMark points I get versus another computer, I care about real-world, usable performance. And a 3DMark score doesn't give me that, ever, at all. A 3DMark score would never help me decide which product to buy, and after-the-fact 3DMark scores tell me nothing about what I've already bought other than it can get a high 3DMark score. Don't care. I get more useful information from running Doom 2.

But hey if 3DMark points are your thing.. :cool: :cool: :cool:
 
griff30 said:
It means marginal clock increases on the initial batches, until it matures.
Click to expand...

i would agree . However Pascal is a totally different architecture so we should a big improvement over the 980ti of at least 30% minimum for the first flagship card.
 
Drakensoul said:
Well since I don't overclock that's gone. And the rest is pointless because it gives me no real world information. I don't really care about how many 3DMark points I get versus another computer, I care about real-world, usable performance. And a 3DMark score doesn't give me that, ever, at all. A 3DMark score would never help me decide which product to buy, and after-the-fact 3DMark scores tell me nothing about what I've already bought other than it can get a high 3DMark score. Don't care. I get more useful information from running Doom 2.

But hey if 3DMark points are your thing.. :cool: :cool: :cool:
Click to expand...

Can you tell me on this graph which card is the slowest and which card is the fastest and why?

index.php
 
griff30 said:
It means marginal clock increases on the initial batches, until it matures.
Click to expand...
Even if true, what the heck does that have to do with anything? Dropping to 16 nm would allow a massive increase in cores/tmus/rops. And going to HBM 2 will be 3 times the memory bandwidth of Titan X/980 Ti. Plus that HBM 2 frees up a lot more power for even more cores/tmus/rops on top of that. And that is before any architectural improvements are even factored in. But yeah I guess if the clock speeds dont increase all that much then its not an improvement...:rolleyes:
 
You must log in or register to reply here.
Back
Top