• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Learning the Basic Terminology??

theMuppet

n00b
Joined
Jan 21, 2013
Messages
8
Hey Guys,

I am fairly new to the entire hardware world, but I finally built a computer from scratch and basically fell in love with this hobby! Love seeing the temperatures fluctuate, trying to make the machine cooler etc..

GPU's are probably the component that excites me the most, and I was wondering if someone could basically send me a link to a thread/site that has all the terminology followed by a good explanation.. Example (SLI, AA, etc..)

Appreciate the help and looking forward to being part of the community!
 
Hey Guys,

I am fairly new to the entire hardware world, but I finally built a computer from scratch and basically fell in love with this hobby! Love seeing the temperatures fluctuate, trying to make the machine cooler etc..

GPU's are probably the component that excites me the most, and I was wondering if someone could basically send me a link to a thread/site that has all the terminology followed by a good explanation.. Example (SLI, AA, etc..)

Appreciate the help and looking forward to being part of the community!

Here ya go:
http://www.tweakguides.com/Graphics_1.html

SLI just means 2 or more Nvidia GPUs in the same machine. Max of 4 cards and they MUST be the same type, ie all 680s or all 670s. AMD calls it CrossfireX
 
...
SLI just means 2 or more Nvidia GPUs in the same machine. Max of 4 cards and they MUST be the same type, ie all 680s or all 670s. AMD calls it CrossfireX

Technically, it doesn't just mean more than one GPU in a machine - it's about aggregating the processing power of the cards. I'm sure you probably know that, but since the OP is trying to learn about it, I think it's appropriate to be pedantic :p
 
Awesome, thanks guys!! Appreciate it

I was looking at video cards and ended up going with a EVGA 670 FTW 2GB card.. I am happy with it but noticed that it does run a little hot when playing games like Crysis 2 while the GPU is OC.. (read a great guide on another forum that helped me out). So I bought two Coolmaster 200mm fan's and it did improve a little (Still cant overclock really crazy because it ends up reaching around 69C and hitting 70-71C

So i was reading around and a lot of people said that reference cards have a crappier cooling system (fan etc..), something a long those lines?

So was wondering what the difference between those cards are?
 
70C is nothing to worry about for a GPU. You've got plenty of headroom before it's going to cause any damage or throttle performance.

"The difference between those cards" is a bit of a vague question - there are a range of different designs out there that will perform differently. Yes, the standard cooler is not the absolute best performance design, but it's been tested by the manufacturer to ensure that it provides enough cooling to do the job.
You could fit a third-party cooler to give you slightly better cooling, but bear in mind that it might void your warranty.
 
Here is wikipedia's description :)

Scalable Link Interface - Wikipedia, the free encyclopedia
http://en.wikipedia.org/wiki/Scalable_Link_Interface


Scalable Link Interface (SLI) is a brand name for a multi- GPU solution developed by NVIDIA for linking two or more video cards together to produce a single output.

Actually SLi was implemented by 3Dfx with the voodoo cards.
Nvidia bought the rights.....and improved it.
 
Actually SLi was implemented by 3Dfx with the voodoo cards.
Nvidia bought the rights.....and improved it.

They didn't buy the rights, they bought out 3Dfx. With that came Glide, amongst a few other technologies like SLI.
 
Last edited:
This info from wikipedia says it all:

used by 3dfx under the full name Scan-Line Interleave, which was introduced to the consumer market in 1998 and used in the Voodoo2 line of video cards. After buying out 3dfx, NVIDIA acquired the technology but did not use it. NVIDIA later reintroduced the SLI name in 2004 and intended for it to be used in modern computer systems based on the PCI Express (PCIe) bus; however, the technology behind the name SLI has changed dramatically
 
They didn't buy the rights, they bought out 3Dfx. With that came Glide, amongst a few other technologies like SLI.

3Dfx SLI and Nvidia SLI are not the same thing and they actually do not even stand for the same thing. The common acronym was mainly for marketing reasons.

http://www.anandtech.com/show/1367/3

This info from wikipedia says it all:

used by 3dfx under the full name Scan-Line Interleave, which was introduced to the consumer market in 1998 and used in the Voodoo2 line of video cards. After buying out 3dfx, NVIDIA acquired the technology but did not use it. NVIDIA later reintroduced the SLI name in 2004 and intended for it to be used in modern computer systems based on the PCI Express (PCIe) bus; however, the technology behind the name SLI has changed dramatically

You guys are picking nits.
3Dfx came up with the scan-line interface to harness the power of two GPUs.
nvidia bought the company....and in it's inclusion was the "right" to use this technology, although they dramatically improved on it's function, obviously now known as Scalable-Link Interface.......
but no matter......3Dfx thought of it first, and that's what i was trying to point out.

Just to make a historical point about what SLi is all about......not to mention Alienware threatened to make SLi a reality several years before nvidia launched the technology during the 6800 series of GPUs. Alienware's implimentation was total vaporware however.
 
I read that the 600 series cards begin to throttle back the boost clock at 70C

This is true, at 70c the boost clock will be lowered. 80c as well.

I haven't reached 90c to test that though ;)

I presume it's at 10c intervals after 70c. That said, that there are not that many intervals of 10c after 70c before you hit the thermal limit. :p
 
Awesome, thanks guys!! Appreciate it

I was looking at video cards and ended up going with a EVGA 670 FTW 2GB card.. I am happy with it but noticed that it does run a little hot when playing games like Crysis 2 while the GPU is OC.. (read a great guide on another forum that helped me out). So I bought two Coolmaster 200mm fan's and it did improve a little (Still cant overclock really crazy because it ends up reaching around 69C and hitting 70-71C

So i was reading around and a lot of people said that reference cards have a crappier cooling system (fan etc..), something a long those lines?

So was wondering what the difference between those cards are?

The "reference" card is generally the stock specification of the GPU and cooling solution recommended by the designer of the chip and circuit board, in this case nvidia.
You are correct, the stock GPU has a generally inferior cooling solution versus a build partner's solution....generally speaking.
Most of the partners embellish the standard card to stand out and of course sell more than the next guy.
The non reference cards usually have a better cooling solution that keeps the GPU cooler and is much quieter than the reference. These cards are often factory overclocked as well.

For nvidia you are looking at offerings from ASUS, MSI, Galaxy, Gigabyte and EVGA.

It's rare to see a non-reference cooler on an evga card. To the other extreme ASUS puts on some giant coolers,usually under the DC2 label. MSI likewise in the Twin Frozr branded cards.
As a blanket statement most of the non-reference cards will run 5-10C cooler and are much, much more quiet.

I have owned a number of TwinFrozr cards. They are usually good overclockers and very cool and quiet.

That said, the truely [H]ard way to cool your GPUs is with water.:D
Once you get wet, you never go back.
 
The "reference" card is generally the stock specification of the GPU and cooling solution recommended by the designer of the chip and circuit board, in this case nvidia.For nvidia you are looking at offerings from ASUS, MSI, Galaxy, Gigabyte and EVGA.

It's rare to see a non-reference cooler on an evga card.

Their Hydro Copper series are non-reference coolers of course, are you saying the EVGA FTW and Classified cards have reference coolers because I think those are customized.
 
Their Hydro Copper series are non-reference coolers of course, are you saying the EVGA FTW and Classified cards have reference coolers because I think those are customized.

Hydros are watercoolers.....not air.

I dont think the FTW or the Classified have anything better than the reference. They are still squirrel cage blowers.....but the heatsink might be beefier ......IDK.
 
I dont think the FTW or the Classified have anything better than the reference. They are still squirrel cage blowers.....but the heatsink might be beefier ......IDK.

Why do people love those cards so much? Due to the boost clock? EVGA doesn't even have a lifetime warranty anymore.
 
70C is nothing to worry about for a GPU. You've got plenty of headroom before it's going to cause any damage or throttle performance.

So if the card hits 70C and stays around 68-72C, will this damage the card? I've been going through a lot of information on different forums and they all say that the 670 starts throttling at 70C (noticed this myself by monitoring Precision X) When it decreases the boost like that when hit hits 70C, is that dangerous?

That said, the truely [H]ard way to cool your GPUs is with water.:D
Once you get wet, you never go back.

I've looked at water cooling and all I found was CPU coolers and couldn't find a GPU water cooling system... Not sure if i am approaching this incorrectly? Is there specific coolers for a GPU or is it basically an entire system that will go throughout the machine??

Why do people love those cards so much? Due to the boost clock? EVGA doesn't even have a lifetime warranty anymore.

I personally picked it as the brand is known and read that their customer service is the best... Sorta newbie! I would have chosen a different card with a better cooling system/fan but what is done is done..

Next computer I build in a few years will be awesome as I will learn everything from you guys!
 
So if the card hits 70C and stays around 68-72C, will this damage the card? I've been going through a lot of information on different forums and they all say that the 670 starts throttling at 70C (noticed this myself by monitoring Precision X) When it decreases the boost like that when hit hits 70C, is that dangerous?

No, Nvidia's specs say 97C is max temp
http://www./hardware/desktop-gpus/geforce-gtx-670/specifications
I've looked at water cooling and all I found was CPU coolers and couldn't find a GPU water cooling system... Not sure if i am approaching this incorrectly? Is there specific coolers for a GPU or is it basically an entire system that will go throughout the machine??

No, the water cooler attaches to the GPU card. You can buy them made that way already, like this card

http://www.evga.com/Products/Product.aspx?pn=02G-P4-2689-KR

Or you can buy a water cooler that you can attach to your card like this one for the 680

http://www.evga.com/Products/Product.aspx?pn=400-CU-G680-B1

I don't water cool so someone else can tell you what water blocks will fit your card.
 
You guys are picking nits.
3Dfx came up with the scan-line interface to harness the power of two GPUs.
nvidia bought the company....and in it's inclusion was the "right" to use this technology, although they dramatically improved on it's function, obviously now known as Scalable-Link Interface.......
but no matter......3Dfx thought of it first, and that's what i was trying to point out.

Just to make a historical point about what SLi is all about......not to mention Alienware threatened to make SLi a reality several years before nvidia launched the technology during the 6800 series of GPUs. Alienware's implimentation was total vaporware however.

I'm not trying to nitpick but to point out that the actual implementation of Nvidia's SLI has no relation to 3Dfx's. The only commonality is the shared acronym and concept of using multiple video cards for rendering. Nvidia SLI is only a continuation and improvement of 3Dfx SLI in the same way AMD Crossfire is.
 
SLI = two single-GPU Nvidia video cards
Tri-SLI = three single-GPU Nvidia video cards
4-Way SLI = four single-GPU Nvidia video cards
Quad-SLI = two dual-GPU Nvidia video cards

If you are not bottlenecked by the CPU and are playing a game that is well optimized for SLI, two GPUs can approach being twice as fast for gaming as a single GPU, but three or more GPUs do not scale well; you can expect no more than 25 to 35 percent more gaming performance from a third GPU, and no more than 20 percent from a fourth GPU.
 
I wouldn't worry about cooling your gpu that much, I've found a slight adjustment in the fan control using evga precision x works fine.
 
As far as cooling goes, look at MSI Afterburner or the precision X. You can make a custom fan profile to lower the temps. But the higher the fan the more noise you'll get.
To watercool you'll need a custom loop,to include a pump,reservoir, radiator, and a gpu block especially made for the gpu.
Id advise reading some in the watercooling subforum.
There are no kits made for gpu cooling like you find for cpu cooling, such as the various Corsair kits.
 
Took a look at some of those kits and man, their not cheap.. Would rather but a few more hundred dollars and get a way better GPU in 2 years or so when this one gets "old".

I've created a custom fan profile and the fan usually runs at 80% and it stabilizes at around 65-68C on current GPU settings. This is basically what I did:

- I've increased the Power Target to 140%
- I've increased the Voltage to the "Max" available
- GPU Clock Offset left at 0%
- Mem Clock Offset left at 0%

After doing that, the GPU Clock went to around 1150 from 1050 or so.. and this is were it is stable at 65-68C... When I increase the GPU/Mem Clock Offset, it starts hitting 71-72%...

Even if its a small increment (5Mhz +) or a very large one (90Mhz +), it stays around 71-72C.

Is this normal behaviour?
 
I read that the 600 series cards begin to throttle back the boost clock at 70C

Ah right, I stand corrected. In that case it would be preferable to lower your temps slightly to avoid triggering the throttle, since that will dent performance.
 
Took a look at some of those kits and man, their not cheap.. Would rather but a few more hundred dollars and get a way better GPU in 2 years or so when this one gets "old".

I've created a custom fan profile and the fan usually runs at 80% and it stabilizes at around 65-68C on current GPU settings. This is basically what I did:

- I've increased the Power Target to 140%
- I've increased the Voltage to the "Max" available
- GPU Clock Offset left at 0%
- Mem Clock Offset left at 0%

After doing that, the GPU Clock went to around 1150 from 1050 or so.. and this is were it is stable at 65-68C... When I increase the GPU/Mem Clock Offset, it starts hitting 71-72%...

Even if its a small increment (5Mhz +) or a very large one (90Mhz +), it stays around 71-72C.

Is this normal behaviour?

Yes. Overclocking creates heat.
That's the benefit of water and stout copper blocks.
 
Thanks guys, appreciate all the answers!! Been reading the tweakguides that was linked above, interesting as hell!

Crazy how much calculations a GPU does to create a single frame and place it into the frame buffer.. and some game run at crazy frame rates like 100FPS

Mind boggling
 
OP,

There are tons of articles online on terminology but if you want what I call a quick " on the crapper" read hit wikipedia and suck in the knowledge. Lots of good stuff.
 
Back
Top