I did it after ~15yrs of only buying team RED GPU

[xkm1948]

Everything started when I had my first modern GPU, Radeon 9700Pro. Before that I have only used Intel's chipset built-in ones or even older, S3 Trio64 cards. Needless to say that GPU got me hooked onto the Radeon team. Ever since then I have been buying exclusively Radeon GPU for my desktop build. X1950XT, 3870, 4870, 5870, 6870 all the way to FuryX.

It has been a long way here. My just retired GPU has been the FuryX. It is an interesting card but the performance as well as general computing capability can no longer suit my needs. I have mixed feelings with the FuryX. There have been a lot of small quirks working with this card. Only one major problem was the AIO pump leaking and I had to pay out of pocket for Sapphire to fix it. However all these years with Radeon makes me feel like that this current RTG is no longer the ATi it once was. After some serious consideration I went ahead with green team this time. The FuryX is not going anywhere. It will be working again in another build I am about to put together.

So which GPU am I getting then?

The EVGA 2080Ti

Here is 2080Ti saying farewell to the old dog FuryX

My computer is mainly for bioinformatics. I do some VR development on the side. Having a CUDA capable card will greatly improve my work efficiency. The amazing performance in VR will also be a great addition.

Specifically I need the Tensor cores for genomics/transcriptomics work. I don't have money to afford a Quadro or TitanV so a gaming GPU with Tensor cores is my best bet. Just in case you are interested in how Tensore cores benefit Genomics, take a look at these video:

As for what I am working on VR development, here is a platform I am working on for protein interactions:

From my brief experience with the 2080Ti I am very pleased with its performance. Well only gaming part for now as I have not updated my Linux to incorporate latest CUDA toolkits. The GPU is quiet and powerful. Most importantly it draws even less power than my FuryX!



Feels weird installing Nvidia driver on my work PC

Still learning how to work with the OC software. Scanner mode looks cool, but I have 0 idea what that +142 means

I guess the grass is greener on the other side.

I wish RTG will be more competitive 3 yrs from now when I will probably be upgrading again

 

[Toothless]

Welcome to the green team.

 

[chaosmassive]

welcome to dark side,err.... I mean green side !

 

[dj-electric]

Welcome to no team, but the ability to fairly judge a product out of experience.
Currently testing and using my own RTX 2080 Ti, not only in gaming but for machine learning. It seem to hold an incredible potential compered to pascal

 

[hat]

Congrats on your purchase. Hope it serves you well!

 

[notb]

You're coding yourself or just using CUDA-optimized software?
Either way, you'll love it.

The card is just way out of my budget. Looking forward to 2070 benchmarks...

Currently testing and using my own RTX 2080 Ti, not only in gaming but for machine learning. It seem to hold an incredible potential compered to pascal

I've talked with someone doing ML who had access to a pre-production RTX. He had some early results. I'm convinced buying Tensor-less GPU is pointless at this point for anything outside of gaming.
I still think Nvidia should reconsider releasing a cheaper card with Tensor - especially for laptops, where these chips are so often use for software acceleration rather than just gaming.

 

[Vayra86]

Enjoy the card!

 

[JalleR]

i did this 2 years ago R9 290(x) -> 1080 and i was blown away, (i had a 4K screen so it wasn’t really fair for the R9 290(x) ) the only thing you can regret is the Price of the 20ti you has to pay, but if you are fine with that you will be very happy, i was with my 1080 and still is with my 1080TI.. thinking about side grading to a 2080 but only because my wife needs a "new" gfx, her 970 is starting to struggle.

"current RTG is no longer the ATi it once was" Amen to that.

I've talked with someone doing ML who had access to a pre-production RTX. He had some early results. I'm convinced buying Tensor-less GPU is pointless at this point for anything outside of gaming.
I still think Nvidia should reconsider releasing a cheaper card with Tensor - especially for laptops, where these chips are so often use for software acceleration rather than just gaming.

it will be interesting to see what the mobile versions of 2070/2080 will bring, but it should have Tensor cores… but Nvidia have bulled off naming scheams before soooooo...

 

[phill]

Welcome and I hope the card does all you need and more

I've had a variety of cards from either team, I've always felt more compelled to go back to AMD than Nvidia.. More so now sadly with the prices That's been a bit of a put off..

 

[qubit]

Wow, you've gone to the green side with a bang!

 

[xkm1948]

You're coding yourself or just using CUDA-optimized software?
Either way, you'll love it.

The card is just way out of my budget. Looking forward to 2070 benchmarks...


I've talked with someone doing ML who had access to a pre-production RTX. He had some early results. I'm convinced buying Tensor-less GPU is pointless at this point for anything outside of gaming.
I still think Nvidia should reconsider releasing a cheaper card with Tensor - especially for laptops, where these chips are so often use for software acceleration rather than just gaming.

Both. Still learning CUDA now

 

[sepheronx]

Can I have your Fury X?

Oh, and congrats on your new card. Very expensive but when you want the best of the best, right?

 

[qubit]

Please do a comparative benchmark (any benchmark) between the FuryX and the new card. I'm predicting 4-5 times the performance.

What CPU do you have? Your specs say "undead CPU".

 

[xkm1948]

Can I have your Fury X?

Oh, and congrats on your new card. Very expensive but when you want the best of the best, right?

Nope you cannot have it. FuryX is staying in the family. It is going to a MiniITX 8700K gaming build

Please do a comparative benchmark (any benchmark) between the FuryX and the new card. I'm predicting 4-5 times the performance.

What CPU do you have? Your specs say "undead CPU".

About 3~3.5X performance if Superposition Benchmark is any indication. Then again FuryX was VRAM limited in that scenario. For CPU just look at my benchmark screenshot above.

 

[qubit]

Nope you cannot have it. FuryX is staying in the family. It is going to a MiniITX 8700K gaming build



About 3~3.5X performance if Superposition Benchmark is any indication. Then again FuryX was VRAM limited in that scenario

I kept my HD 2900 XT (BIOS flashed pro) for sentimental reasons, too. Pretty useless for anything now, but still looks gorgeous.
Will be interesting to see what you actually get in the benchmark. EVGA top brand, too.

 

[Joss]

FuryX is staying in the family.

Hi xkm, I'm your long lost cousin

 

[sepheronx]

Yeah, I'm also a relative.

 

[droopyRO]

My computer is mainly for bioinformatics. I do some VR development on the side. Having a CUDA capable card will greatly improve my work efficiency. The amazing performance in VR will also be a great addition.

Specifically I need the Tensor cores for genomics/transcriptomics work. I don't have money to afford a Quadro or TitanV so a gaming GPU with Tensor cores is my best bet. Just in case you are interested in how Tensore cores benefit Genomics, take a look at these video:

Yes, i use my 980 Ti for those exact things also.
/sarcasm

 

[king of swag187]

Congrats! Not on just going team green, but going all out!! If you're interested in selling that Fury, I'm interested in Crossfire surround as of now......

 

[cucker tarlson]

brand loyalty is silly from a consumer's standpoint.

 

[27MaD]

YOU TRAITOR! , i mean.... , oh that's good , welcome to the green team.

 

[John Naylor]

I pride myself on being and try and educate our users to be what I call "hardware whores". Brand loyalty will bit ya on the ass every time. he he has the best numbers captures my dollars. I generally advise users to avoid the 1st stepping of GFX cards as proces are ridiculousyly high, there are often oopsies, and I want to see who did the best job. And, at this point in time, haven't seen many reviews ... at least not ones worth reading that do a nice teardown. EVGA especially has been a bit tight with their provision of samples to reviewers who do complete teardowns ... understandably so.

-One-third of the original 970 SC's cooler actually "missed" the GPU. This was corrected in later steppings, hoever, there were no physical changes tot he PCB which was significantly less robust the the competitions AIB cards
-All of the early 1060 - 1080 SC and FTW cards didn't have the necessary thermal pads, and by saving a few cents, we saw a wave of stories about cards going up in smoke. There was a recall, you could either send back for a repair or they would send you a kit containing thermal pads.
-MSi flubbed their early 9xx series by using an adhesive tape meant to imobilize the fans during shipping. It turned out that the tape supplier went quite a bit over spec on the adhesive and many users damaged their fans when removing the tape.
-Early Asus cards fan feature link to sync case fanbs with GPU fans is broken on earl 20xx series cards

All of these were / will be fixed with later stepping cards. One other item ... since the 5xx series, the SC line has almost always used a reference PCB / VRM / Power delivery system... again, competitors AIB cards in the same price niche almost always use custom PCBs w/ beefed up VRMs / Power delivery. Of course, if not overclocking, this will matter little.

The data collected from the TPU tests is summarized in the table below. My take-aways from that were a) Asus did a great job with the cooler, tho one has no reason to care really if it's < 75C and b) MSi really cut the power loose allowing the game to pull large amounts of power when needed ... giving it a 3.4% advantage in performance (used W3 at 1440p) Overclocked heaven performance advantage over the Asus was about 0.7%

Other than that, haven't been able to find many reviews that do beyond the rewritten promo stuff. Guru3D and TPU have the only reviews woth reading that I have found so far and so far none have tackled the EVGA model. If you find any, or can pull any data from what you have rec'd, let us know. be noice to know what the better options are before the tariff's jump from 10 to 30%.

 

[xkm1948]

@qubit here is the comparison you want

About 2.9X if you look at just graphics performace

 

[king of swag187]

@qubit here is the comparison you want

About 2.9X if you look at just graphics performace

View attachment 108008

*sniffs* its beautiful

 

[lexluthermiester]

brand loyalty is silly from a consumer's standpoint.

Exactly! Be loyal to yourself first! Choose the products that do what you need them to do. In this case the buyer had a specific need that the 2080ti delivers.