In fact some of the Factory OC 780 surpress GTX Titan's performance.
GTX 780 is the second fastest single gpu - so if you can go with it without any second thought.
Then i am going for the OC versions.
Ok Guys?
Ok TOP
Ok, so I just called SMC international and they said ASUS DC2 version will not be available for around 2 months due to over demand.
Gigabyte OC version they have for 47K ,..... god Too much Pricey:eeksign:
What to do Guys????
Do they have the asus dc2 version in stock?
If they are quoting 47k including tax for the gigabyte windforce oc version of the 780, you can go ahead and buy it.
Else if you don't get any non-reference 780, go for a reference 780,the titan cooler looks better than any non reference cooler and also keeps the temps down.A single 780 is enough for your needs.
Look for upgrading next when maxwell or volcanic island gpu's come out in q3 or q4 2014.
Do they have the asus dc2 version in stock?
If they are quoting 47k including tax for the gigabyte windforce oc version of the 780, you can go ahead and buy it.
Else if you don't get any non-reference 780, go for a reference 780,the titan cooler looks better than any non reference cooler and also keeps the temps down.A single 780 is enough for your needs.
Look for upgrading next when maxwell or volcanic island gpu's come out in q3 or q4 2014.
I think Gigabyte GTX 780 OC @ 47K is a real good choice. Most of the reviews have shown that when overclocked properly, GTX 780 can suppress the gaming performance of GTX Titan in mot of games. Now Arnab, tell me one thing; are you planning to SLI to GTX 780 in future and getting the 1st one for now? If not then you can get two GTX 760 which is priced similarly lile a GTX 660 (around 20 to 21K) and SLI them. Their performance is way over single GTX 780 and comes almost 8K cheaper. Nvidia also has better driver support for SLI
I would suggest to wait for AMD's Curacao and Hainan cards. Curacao is supposed to be a competitor of 780. If AMD can price them competitively as well as perform above expectations, then bagging one of them will be a very good idea. I suggest to wait a while. 47k is indeed a big amount and will hurt a lot if curacao will perform like 780 costing 35k.
Btw check the latest 780 sli review
*hardocp.com/article/2013/08/06/nvidia_geforce_gtx_780_sli_video_card_review#.UgH4o5JTAzI
They give a much better experience than 3 7970 ghz tri-fire. Radeons are indeed a disaster currently at eyefinity resolutions.
No wonder our forum member d3p was troubled so much.
In GCN architecture, AMD has designed their memory modules not for just higher bandwidth but for superior computing power too. They are designed in such a manner that programmers will find it easy to program with GPU computing based languages. In fact wider memory bus does have higher performance impact when coupled with wider memory bus. In GCN, AMD actually brought best of the both world, not just gaming performance or Compute performance.If you look at it closely, you will find it is more like a CPU buss of a many core CPU architecture where multiple cluster of memory bus can access Video Ram simultaneously and assign threads more efficiently to the different compute engines fo better efficiency and resource utilization. Also the design is suitable for X86 type memory management techniques which helps the programmers greatly to optimize the code blocks.
In Kepler design, nVidia almost completely omits the GPU computing part and focused on pre gaming performance. Their memory bus is more like fetching as much data as possible in every clock cycle to feed the CUDA cores. Now in gaming, the design is the best case as here GPU needs to perform very limited number of operations over millions of pixels and coordinates. But in case of general programming, not all the time, you have a huge chunk of data available for a single operation. Here the data needs be divided into smaller groups based on the type of instructions which are needed to be operated over them. in that case, more number of memory bus clusters (384 bit bus is divided into 6 chunks of memory controller of 64 bit width, each handling 512 MB of memory Block), can independently fetch data from their respective memory locations simultaneously and assign them to different memory compute engines. It also makes Thread scheduling more easier.
waiting for the Upcoming AMD GPUs .
Armab, you acn go for this instead. Makes much:
Radeon HD 7990 Price Cut to $799 After Catalyst 13.8 Release
^^ Arnab i think 7990 isn't available in online shops in India now. The stocks must have exhausted.
Wait for the new prices to reflect here in India. Contact your nearest dealer like MD computers or Vedant.
Contact Cilus if possible.
the upcoming flagship gpu from AMD is supposed to give better performance than Titan - now if this happens the price of Titan is going to fall. Now HD7990 with reduced price looks like a better option but still AMD needs to fix the short comings of FPF [ frame pacing fix ] driver very quickly to make HD7990 more attractive. Now with all these chnages with team red nvidia won't sit silently and may release some new products too after that.
Anyway, so now you have three options :
1. Get the flagship AMD gpu after it's release on this year and CF later by adding one more later.
2. Get 2x Titan or GTX 780 if the prices comes down after the release of of the above.
3. Get a HD7990 with updated price but this may cost you 55-60k and wait for better driver.
BTW, I've a little question : 2x HD7970 GHz in CF or one HD7990 which one is better in terms of heat generation [ runs coolers ], consumes less power and is more durable [ read gpu lasts long ] ??