GeForce GTX 650 Ti Boost Review: A Budget-Oriented GK106-Based Boss

Page 6 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.


Radeon 7770 and GTX 650 Ti were competitors until the Radeon 7790 showed up. The GTC 650 Ti Boost is a competitor for the Radeon 7850 and merely has a somewhat lower reported price. The Radeon 7790 is still not its competition because it is specifically competition for the GTX 650 Ti now.

That the 7790 was meant to be competition for the 650 Ti is totally irrelevant in the context that we were in because the 7790 is still not here and we didn't even know about it until recently. Before it, the only competition for the 650 Ti was the Radeon 7770.

The 7790's MSRP is not nor was below that of the 650 Ti.

Besides, about all of this middle of the road crap with the 650 Ti between the 7770 and the 7850/650 Ti Boost... Do keep in mind that in decent quality settings and such, the 7770 is literally right behind the 650 Ti. The 7790 makes a better case for being in the middle because it's actually in the middle instead of only a little bit above the 7770 and far below the 7850.
 

Evolution2001

Distinguished
Oct 16, 2007
108
0
18,680
0
I don't know that I'm completely onboard with these types of news articles. They are good and fun to read, but until the cards are actually in retail channels, I think it's unfair to AMD to repeat whatever nVidia has stated on a phone call.
If memory serves me correctly, nVidia has been known to spout off about new hardware and expected pricing, only to see the product launch date arrive and..."but it's only available in (insert country name) and there are only 3 units available. And the price has been marked up due to limited availability."

So when we get info 'leaked' like this, many of us on the verge of purchasing an AMD card suddenly decide to wait. Granted, waiting always guarantees a lower overall price regardless of which brand you choose, but it seems unfair to AMD.
This scenario can apply to either mfg, and I have no allegiance to either of them, but it seems like this scenario happens more with nVidia than AMD.

On the other hand...hey, it's guerrilla warfare marketing. And if nVidia is better at it, then so be it! If it means better prices for the consumer, that's good. But not-so-good in the long run (for consumers) if nVidia squeezes everyone out because of their aggressive ("questionable"?) marketing tactics. Then as the only major player, their prices are not likely to benefit the consumer.

Meh...whatever. As you were... :)
 

ericjohn004

Honorable
Oct 26, 2012
651
0
11,010
16
The 7790 needs to drop to 130$. The 7850 1GB needs to drop to 145$ to compete. And I'm not talking about after rebates and games. This needs to be their starting price. AMD's starting price needs to be below what Nvidia is asking because perception is everything.

People perceive Nvidia to be quality and they perceive AMD to be second best. Therefore AMD needs to be cheaper and better performing to match the "quality" Nvidia is putting out.

And maybe Nvidia is higher quality, if you read the review on SLI and Crossfire 660Ti's vs 7870's the 660Ti has practically no tardy/dropped frames while the 7870 has quite a few. The 660Ti's graph is perfect while the 7870's graph is jumping up and down.

So maybe there is something to this perception that people have. Not all people have this perception. But clearly, from the sales numbers highly favoring Nvidia while the price/performance seems to favor AMD. So it doesn't make sense. And now that price/performance is favoring Nvidia, AMD better watch out.

I can't see anyone buying the 7790 over a 650Ti w/Boost 1GB.
 

amd damaging the brand new lineup's user perception by neglecting to support cost amd and nvidia took advantage of it (and gained). nvidia's timing and execution had also been better than amd's.
nvidia's switchable gfx tech matured better than amd's, which led to oems demand more of nvidia's gpus while amd... well, they did whatever they did. i don't remember what amd even had at that time. but i do remember optimus' launch and p.r.
in gcn launch's case, the cards had gone without amd's attention long enough for people to withold purchase or fall back on older and/or less performing nvidia cards that provided better overall experience.. it was not for long, because kepler launched, but it certainly happened. a lot of those people later went straight for gtx 680, 670 when those launched. i think, for example, this may be one of the reasons why gtx670 and 680 rank to high on steam's video card rankings in hwsurvey. in amd's defense, may be more nvidia users use steam than amd users. :p

i agree about pitcairn's pricing. but nvidia's timing with deals was better.

quality has almost nothing to do with sales. if it had, no one woulda bought gtx550ti. :) my point was that amd put in the hardware, hyped it, then left it dark for months. even reviewers were vocal. that hurts image.

Lawl, it's quite relevant to my point - how nvidia sells more cards (i.e. why people want to buy nvidia cards more than amd/how nvidia makes more money) than amd. you picked up only on "drivers". it's okay if you didn't understand it. darn semantics, eh? :D
nvidia making money by getting customers to buy one pro card with limited compute perf. and then get them to buy another with high compute perf. is just good business. all these despite cuda being locked and proprietary. :) otherwise amd and nvidia's position would be much different in pro/hpc market and nvidia wouldn't sell as many cards - gaming or pro. in mainstream situation it'd be like using an nvidia card a for physx, even when one uses an amd card as main gfx card.

you're right, finally. :)
 

slomo4sho

Distinguished
Oct 30, 2008
3,199
0
21,160
171


How am I quoted for something I never said :ouch:
 

Justarius

Distinguished
Nov 1, 2007
4
0
18,510
0
I don't understand why this review is comparing the 2GB 650 Boost to the 1GB 7850. Both cards come in 1GB and 2GB variants, why aren't we comparing apples to apples? It's not like the 2GB 7850 wasn't readily available, if anything it's the 1GB that is hard to find... It first I assumed this was why Anandtech's review scored higher for the 7850 but I see some of the other sites did a proper comparison of the two and still score the 650 Boost higher. Still, seems sloppy Tom's would use unbalanced memory and then make such a big deal over the comparison. None of the other reviews did this.
 

Duke Nucome

Honorable
Mar 6, 2013
238
0
10,690
3



It's because the 1GB 7850 is the same price as the 2GB 650ti boost. That said these card don't even have the core power to warrant more than 1GB Vram.
 

ingtar33

Illustrious


bingo. the 7870 is the lowest end gpu to warrent 2gbs of vram. neither the 7850 nor the 650ti boost will saturate 1gb... unless they're in xfire/SLi, but that's about the only time they might saturate 2gb. Get a 2gb version if you plan to xfire/sli low end gpus in the future. otherwise skip them.
 

Isaiah4110

Distinguished
Jan 12, 2012
603
0
19,010
17
I'm pretty sure common knowledge states that the 2GB version of the 7850 gives much better performance at 1920x1080 than the 1GB version, especially if you want to use any sort of AA...
 

Duke Nucome

Honorable
Mar 6, 2013
238
0
10,690
3


http://www.bit-tech.net/hardware/graphics/2012/11/05/ms...
In Battlefield 3 at 1,920 x 1,080, the MSI HD 7850 1GB was actually able to outperform the 2GB version thanks to its factory overclock, albeit by a single frame per second on the minimum frame rate. This suggests that the 2GB's extra 1GB of GDDR5 is all but useless at this resolution.

http://www.legitreviews.com/article/2048/13/
When it comes to gaming performance the XFX Radeon HD 7850 1GB Core Edition did better than expected with a screen resolution of 1920x1080! We were able to easily play all the current game titles at 1080p with the image quality settings cranked up! There were a few game titles like Skyrim where we felt a minor hiccup here or there when the frame buffer would fill up, but they were few and far between.


http://www.techradar.com/reviews/pc-mac/pc-components/graphics-cards/msi-hd-7850-1gb-1129390/review
"After all, the 2GB frame buffer of the stock HD 7850 must be partly responsible for its impressive showing at hi-res, right? And that's putting post-processing such as anti-aliasing in the mix.
How wrong we were. It seems that the bottleneck to the Pitcairn Pro's performance is not the memory at all - the GPU itself bottlenecks performance well before the memory gets anywhere near making things crawl.
What does that mean though? Well, it means the 1GB version of the HD 7850, especially the slightly overclocked version supplied by MSI, is more than able to keep pace with the stock 2GB version of the Pitcairn Pro card."
 

Duke Nucome

Honorable
Mar 6, 2013
238
0
10,690
3


It often is when the hype machine gets all revved up and people start to spread somewhat inaccurate information. Truth is Vram is more of a marketing tool used to sell more product than anything. The only place that I have noticed where more Vram makes a difference is triple monitors.
 

ingtar33

Illustrious

bingo

multiple monitors and huge resolutions will saturate 2gbs... on a single monitor at 1080p you only need 1gb.

I think gaming across 3 monitors at super high res might require 3gbs... nothing will really need more then that.
 

Slavens

Distinguished
Jan 11, 2007
15
0
18,510
0
I think I missed something here. This isn't an attempt to troll, I'm trying to figure out how the benchmarks lead to the 650 ti Boost besting the 7850.

I'm not an 'ATI or NOTHING!' freak. I buy what makes sense at the time.

First, the 7850 is missing from Conclusion, Performance per dollar, so you can't just make a quick direct comparison:
http://www.tomshardware.com/reviews/gaming-graphics-card-review,3107-performance.html

Second, Tom's compares the 2gb 650 ti Boost against the 1GB 7850 and the 7850 bests the ti Boost in three of the five titles

As for the OpenGL, I'd like to see actual openGL games benchmarked. There's more to an openGL title than pure rendering. There's far more to a game than rendering. We've seen the impact of ambient occlusion, 'god-rays', and extreme anti-aliasing setting on the raw framerate of video cards.

I also have to question the use of Autocad as a benchmark, mainly because I feel that it places too heavy of an emphasis on pure rendering. While synthetic benchmarks are always part of of a good review, we benchmark performance in specific games using specific video configurations because we know that these provide two important bits of information:
- a 'real-world' expectation of performance of a given product
- a point of comparison between competing products in a meaningful manner, across price points and time (as new cards are released)
If we're going to do OpenGL/CL benchmarks in Bang-for-the-buck gaming cards, I'd like to duplicate the D3D efforts. If we benchmark the D3D systems synthetically, do the same for OpenGL. If we benchmark the D3D with games, please use games for OpenGL.

I also have to express disappointment in the lack of discussion over the memory bandwidth. If you look through reviews of video cards that have a similar GPU but varying memory buses, you'll see that cards with higher memory bus widths see smaller framerate drops as video settings are increased. Similarly, cards with high memory bus widths typically see longer service lifes - take the Radeon x850 series for example. I'd give an nVidia example, but I'm not as familiar with their cards.


My problem is that the 1GB Radeon 7850 bested the 2GB 650 ti Boost it in 3 out of 5 games with half the RAM and has a 75% wider memory bus. I don't see how the Boost 'wins'.

Further, at the moment, the 1gb versions of these two cards are both $149 on NE and the
2GB 7850 is $10 cheaper than the 2GB Boost at $159 and $169 respectively. Given the $10-$20 difference between the 1GB and 2GB versions of the cards, I'd skip the 1GB versions entirely.
 

Slavens

Distinguished
Jan 11, 2007
15
0
18,510
0
I think I missed something here. This isn't an attempt to troll, I'm trying to figure out how the benchmarks lead to the 650 ti Boost besting the 7850.

I'm not an 'ATI or NOTHING!' freak. I buy what makes sense at the time.

First, the 7850 is missing from Conclusion, Performance per dollar, so you can't just make a quick direct comparison:
http://www.tomshardware.com/reviews/gaming-graphics-card-review,3107-performance.html

Second, Tom's compares the 2gb 650 ti Boost against the 1GB 7850 and the 7850 bests the ti Boost in three of the five titles

As for the OpenGL, I'd like to see actual openGL games benchmarked. There's more to an openGL title than pure rendering. There's far more to a game than rendering. We've seen the impact of ambient occlusion, 'god-rays', and extreme anti-aliasing setting on the raw framerate of video cards.

I also have to question the use of Autocad as a benchmark, mainly because I feel that it places too heavy of an emphasis on pure rendering. While synthetic benchmarks are always part of of a good review, we benchmark performance in specific games using specific video configurations because we know that these provide two important bits of information:
- a 'real-world' expectation of performance of a given product
- a point of comparison between competing products in a meaningful manner, across price points and time (as new cards are released)
If we're going to do OpenGL/CL benchmarks in Bang-for-the-buck gaming cards, I'd like to duplicate the D3D efforts. If we benchmark the D3D systems synthetically, do the same for OpenGL. If we benchmark the D3D with games, please use games for OpenGL.

I also have to express disappointment in the lack of discussion over the memory bandwidth. If you look through reviews of video cards that have a similar GPU but varying memory buses, you'll see that cards with higher memory bus widths see smaller framerate drops as video settings are increased. Similarly, cards with high memory bus widths typically see longer service lifes - take the Radeon x850 series for example. I'd give an nVidia example, but I'm not as familiar with their cards.


My problem is that the 1GB Radeon 7850 bested the 2GB 650 ti Boost it in 3 out of 5 games with half the RAM and has a 75% wider memory bus. I don't see how the Boost 'wins'.

Further, at the moment, the 1gb versions of these two cards are both $149 on NE and the
2GB 7850 is $10 cheaper than the 2GB Boost at $159 and $169 respectively. Given the $10-$20 difference between the 1GB and 2GB versions of the cards, I'd skip the 1GB versions entirely.
 

somebodyspecial

Honorable
Sep 20, 2012
1,459
0
11,310
20


Any proof? IN what? Cuda is still in adobe so basically AMD is handing over some cash to get it in there. No surprise, it's a good move as lots of content comes from that suite.

Proof of rapidly replacing cuda please. I disagree with Toms workloads on opencl also. When there is a way to test cuda (via some other plugin etc) you should be comparing them both to each other, not showing how bad opencl is on NV (well duh, because they already can do X in Cuda so why waste on opencl?). F@H/Bitcoin etc is uncommon. Opencl for crap like blur, photo stuff, vid editing etc is uncommon, people use adobe and cuda. You will finally be able to compare these soon, but they should compare NV running cuda on adobe, vs the same photo or vid job on opencl for AMD. Pretending cuda isn't used like this (adobe) is misleading. It amounts to FINDING a way to do these operations in opencl so AMD can actually attend the party, instead of just using Cuda in adobe like the majority of content people do currently.
 
why everyone assuming Adobe already drop CUDA? for all i know the latest Adobe still have CUDA in them.

http://blogs.nvidia.com/blog/2013/06/19/5-reasons-to-choose-nvidia-gpus-for-stunning-performance-in-adobe-creative-cloud/

OpenCL might be much more known than before but it doens't mean it will kill CUDA just like that. both CUDA and OpenCL have their own advantage over another.
 
Status
Not open for further replies.

ASK THE COMMUNITY

TRENDING THREADS