GTX 650ti Boost VS GTX 660 | STFUandPlay

GTX 650ti Boost VS GTX 660

By Torrence Davis Thu, April 11, 2013 - 9:10:39

Nvidia has really changed the game this year.  First they introduced the GTX 660 to the world.  A sub $200 card that could play games with ultra settings at 1080p or higher.  It was the ultimate mid-level gamers card and it could play just about anything that was thrown at it with little to no sacrifices in performance.  Later Nvidia released the GTX 650, an entry level card that lacked the much needed performance to play games like Battlefield 3 on high settings.  This card was later updated to a 650ti, but still wasn’t an attractive card for gamers.  Nvidia needed something that could compete with AMD’s 7850 card in price and performance, so in came the GTX 650ti Boost.  This new card was an overclocked version of the 650ti that offered all the bells and whistles of its older brother, the GTX 660.  It has the same clock speeds, memory bandwidth and size, lower power draw (134W vs 140W) and a lower price.  It’s almost as if they just dropped a few CUDA Cores on the 660 to make this card.  They are nearly identical.  The real trick here was to see how well the 650ti Boost stacked against the 660 in price and performance.

If you’re a newcomer to PC gaming with a modest budget, you’re probably looking at the 600 series of cards and scratching your head.  You can find a 650ti Boost for as low as $169.  The cheapest 660 I found was $189 and both of these cards are by EVGA.  The cards I reviewed were reference cards and are clocked lower than some of the EVGA cards that are available.  For instance, the Nvidia reference 650ti Boost has a core clock of 980MHz and a boost clock of 1033MHz.  The $179 EVGA version I looked up on Newegg has a core clock of 1072MHZ and a boost clock of 1137MHz.  That’s faster than the card we reviewed and also faster than the reference 660 we used in our benchmarks.  The EVGA 660 base model has a core clock of 1046MHz and a boost clock of 1111MHz in comparison.  Keep in mind that the 660 has more CUDA Cores than the 650ti Boost (960 vs 768).  This makes all the difference in performance between the two cards.  So with a budget in mind, you’re only going to pay $20 more for the 660 if you’re a smart shopper. 

I could throw numbers at you all day, but they mean nothing without a true test of their performance in a gaming environment.  My testing platform is a first gen Intel i7 920, with 6gb of RAM.  It’s a beast of a machine but pales in comparison to some of the newer machines being built today.  While I consider myself savvy with PCs and build my own, I’m no extreme enthusiast.  You’ll never see me running SLI or loading up my motherboard with 24gb of RAM.  My testing machine is what I use every day for HD video editing, RAW photo editing, internet surfing and gaming.  It’s not a bread box setup like you might find at PC Gaming or Tom’s Hardware.  This is as real world as it gets.

650ti Boost vs 660

The games that I chose to test are games that I’m actually playing right now.  I ran benchmarks for Battlefield 3, Tomb Raider, Bioshock Infinite, Call Of Duty Black Ops II and Planetside 2.  I used Fraps to do my benchmarking and while it recorded min, max and average frame rate, I only included the average frame rate in my chart.  Oddly enough Planetside 2 maxed out at 250fps for both cards and I can’t figure out why.  Also, most of the games had a max fps of 63.  These results are obviously scalable. So running a newer AMD or Intel PC will net you far better frame rates.  You can use this chart as a baseline.

Every game I tested was run with Ultra settings at 1080p.  As you can see from the chart, the 660 edged out the the 650ti Boost by about 3-11 fps.  This difference wasn’t noticed at all in any games except for Tomb Raider.  TR just ran a bit smoother on the 660 then it did on the 650ti Boost.  Call Of Duty BLOPS 2 is locked at 60 and doesn’t have an Ultra mode.  So both cards performed pretty much the same with COD.  Planetside 2, Bioshock Infinite and Battlefield 3 differences were minimal, but there is a performance boost with the 660.  Now if you apply these results to an Intel 3770K for example, you’ll probably see a much higher performance ceiling but I don’t think you’d see much more of a difference between the cards.

So does the 650ti Boost give you more bang for your buck?  It really depends on how much bucks you have.  The 660 will perform better with games that push the hardware.  For instance, Tomb Raider is the only game that’s running TressFX.  Obviously the 650ti Boost can’t perform as well with TressFX turned on, but still supports it.  And I’m only assuming that this is what caused the performance difference between the two cards.  To some, an 11fps gain is worth the money.  I really can’t tell you which card you should buy, but if you want to shave a bit off the cost of your system build and are content with 1080p Ultra settings at 35+ FPS, the 650ti Boost is a good choice.  If you want to future proof your system for games that might utilize more special F/X like TressFX, I would spend the extra money and get the 660.  Games like BF4 will most definitely run better on this card on Ultra settings than they would on the 650ti Boost.  Either way, they are both great products and you can’t go wrong either way.

comments powered by Disqus

Member Links

E3 2014 Central

Real Time Web Analytics