MSI GeForce GTX 460 Cyclone 768MB OC Review

[SKYMTL]

There is no denying the fact that NVIDIA’s GTX 460 has been extremely well received and represents exactly what many were waiting for in terms of performance and price. However, every single board partner stumbling over themselves to release these cards, it is getting a bit hard to distinguish one from another. On the day of launch there were two memory capacities (768MB and 1GB), overclocked and custom cooled products galore and they all retailed within $40 of one another. We don’t envy anyone who steps into this field expecting to make a quick decision but we can help point you in direction…hopefully.

MSI’s products and their graphics cards in particular haven’t been reviewed all that much here at Hardware Canucks but the number of offerings they have is simply staggering. On the higher-end front are the lightning-branded products with Twin Frozr heatsinks and what MSI calls “Military Grade Components”. Slightly lower-end cards are available as well sporting various levels of overclocks along with the usual repertoire of reference-based GPUs. All in all, their lineup is more than complete.

While MSI hasn’t released their GTX 460 Lightning editions yet (we assume they’ll be announced shortly), a number of pre-overclocked SKUs have been introduced as OC Editions. Abstaining from the stock heatsink, MSI is using their excellent Cyclone cooler on these products even though the rest of the design is kept at NVIDIA’s original specifications. These Cyclone OC cards are available in both 768MB and 1GB flavors but in this review we will only be taking a look at the lower-end 768MB version.

Unfortunately, MSI doesn’t have the lifetime warranty that some of their competitors market but with three years of coverage, they should have you covered for the time you intend on keeping this card in your system.

The pricing for the GTX 460 Cyclone OC is currently sitting at $210 USD - a mere $10 more than a reference GTX 460. This also puts it at $10 less expensive than EVGA’s own Superclocked Edition which we reviewed a little while back. Let’s see how this thing performs!

 

[SKYMTL]

Say Hello to the GF104

Say Hello to the GF104

When it came to shrinking the GF100 which graces the GTX 480, GTX 470 and GTX 465, NVIDIA looked closer at the architecture and decided to make a few changes to a number of areas. In order to scale back things, certain sacrifices would have to be made if the original GF100 layout was kept as we speculated in our original article. The main issue with the GF100 is a real lack of texture units as you start eliminating Streaming Multiprocessors. So, if linear scaling was kept, NVIDIA have had possible lower-end GPUs with 320 cores or fewer cores but only 40 or fewer texture units. Like it or not texture performance is still one of the cornerstones of modern games and if that trend had continued, NVIDIA may have found it very hard to compete with the HD 5000 series.

One of the primary reasons behind designing the GF104 was the need to lower the thermal and power consumption needs of the Fermi architecture by producing a more compact core. Not only is this easier and less expensive to produce but it also allows NVIDIA to attack certain price points which ATI may have left vacant.

The differences between the GF100 and GF104 layouts start with the Streaming Multiprocessor which houses the CUDA cores, Texture Units, Polymorph Engine, Warp Schedulers, Load / Store units, SFUs and their associated cache hierarchies. Let’s start at the top and make our way down.

Instead of two Dispatch Units each being accessed by their own Warp Schedulers, the GF104 makes use of a 2:1 ratio between the dispatch units and the schedulers while the number of Special Function Units has doubled per SM. As a result, transcendental instruction performance has been increased over the GF100 even though the number of concurrent threads has remained as it was. Otherwise, the Instruction Cache and the Register File size stay the same as GF100.

The main changes to the SM come with the number of CUDA cores as well as the number of texture units each houses. Instead of the usual 32 cores per SM, the GF104 uses a structure which allows for 48 cores along with 16 load / store units and 8 Special Function Units. This in and of itself is quite an eye opener but the real differences are with the number of texture units each Streaming Multiprocessor houses. The GF100 cards have four texture units per SM while the NVIDIA equipped the GF104 with eight TMUs per SM. This can and will lead to a massive increase in texture performance which will benefit older DX10 and DX9 games.

Much like the GF100 layout, the GF104 makes use of four Streaming Multiprocessors (SMs) and their associated Polymorph Engines per GPC along with a common Raster Engine. The only differences are the ones we mentioned above and the result is a GPC with sufficiently more horsepower than the GF100 was able to put forth.

Above is a picture of a full GF104 core and we’re hoping you are paying very close attention to its layout and the number of cores it houses within its two GPCs. In total, there are 384 cores, 64 texture units, 32 ROPs, 512KB of L2 cache and four 64-bit memory controllers. It is quite evident that even though the GF104’s SM structure got a face lift, NVIDIA kept the ROP, L2 cache and memory controller array as is when making the transition from the GF100.

To us it looks like NVIDIA took some of the lessons it learned from the GF100 and put them towards designing a core that is infinitely more adaptable for the sub-$250 market. Not only is the GF104 much more compact than the higher-end silicon (it has 1.95 billion transistors versus the GF100’s 3 billion) but it is supposedly quite a bit more efficient as well. The one thing which could hold it back is the fact that it only has a maximum of eight PolyMorph Engines that are essential for DX11 performance. For example, if you wanted to achieve 384 cores with a GF100, a total of 12 SMs (and 12 PolyMorph Engines) would be needed. Will these eight or fewer engines have a negative impact on the GF104’s DX11 performance in future applications? Only time will tell but for the time being this looks like the perfect graphics processor for the current mid-range market.

 

[SKYMTL]

The GTX 460: GF104 Slimmed Down

The GTX 460: GF104 Slimmed Down

According to NVIDIA, the GTX 460 is meant to compete with ATI’s HD 5830 from a number of different aspects. We have always maintained that ATI’s lineup is a bit weak between the $170 and $250 price brackets and this is exactly where we see this new card hitting. Before we go on, it should also be mentioned that the GTX 460 is compatible with 3D Vision Surround and NVIDIA Surround as well.

The GTX 460 is being offered in two different flavours that are distinguished from one another by their price and memory size: 1GB and 768MB. At face value, the distinguishing qualities between these two cards may be minor but since the L2 cache, ROPs and memory controllers scale in a parallel fashion with one another, the elimination of 256MB of memory causes a bit of a domino effect. So the 768MB card not only ends up with less memory than its bigger brother but also less ROPs, 384KB instead of 512KB of cache and significantly narrower bandwidth as well. Clock speeds, core / texture count and other aspects stay constant between the two cards. What’s even more impressive is the number of texture units equals those on a GTX 470 while TDP (not to be confused with actual power consumption) is actually quite low for a 400-series card but remember that ATI’s figure of 171W for the HD 5830 is based on max board power.

When you look closely at NVIDIA’s lineup, everything looks extremely well defined until you throw the GTX 460 into the mix. The main “problem” we see is the GTX 460 stepping on the toes of the $270 GTX 465 and could make the more expensive card look like a lame duck in terms of pricing, efficiency and even performance. The only area where the new kid on the block looses out is with the number of cores but in general, its paper specifications do make for some impressive reading. This will be great news to those of you who didn’t jump on the bandwagon but it also causes us wonder if the GTX 465 is being pushed out of the market already. There is however the small matter of the GTX 465’s edge when it comes to the PolyMorph Engines which could give it a serious edge in some DX11 applications.

One of the most interesting aspects of the GTX 460 is that it doesn’t actually sport a full GF104 core. Much like the GTX 480, NVIDIA decided to disable a single SM in order to improve yields and (in a roundabout way) ensure this sub-$250 card wouldn’t end up taking a bite out of the whole 400-series lineup. Will we see a 384-core totting, GF104-based card in the future? You never know but NVIDIA is likely keeping this as an ace up their sleeves in case ATI is able to mount a counter-offensive.

Unlike the GF100 cards, the GTX 460 also supports full bitstreaming of HD audio over HDMI. All of the signal processing is done on the card itself without the need for external decoding. This is a huge step forward for those of you who want to use this card in an HTPC environment for decoding Dolby DTS-HD Master Audio and TrueHD tracks.

 

[SKYMTL]

MSI GTX 460 Cyclone Specs / Packaging & Accessories

MSI GTX 460 Cyclone Specifications

MSI equips their Cyclone OC card with a higher core clock but the memory speeds unfortunately stay at reference values. The reasoning behind this lack of memory overclock isn’t quite clear since we have seen in the past that the GTX 460’s GDDR5 is highly adaptable to being pushed. It could be that MSI is leaving room for additional cards in their lineup but it would have been nice to see at least a small bone thrown our way in terms of increased memory performance.

Packaging & Accessories

​

There was once a time when graphics cards manufacturers loved to hike up the size of their boxes in order to get noticed on retail shelves. The result was astronomical shipping costs and a massive waste of cardboard. Luckily, those times are over and companies like MSI have begun shipping their GPUs in more modestly sized boxes. This one gives us all of the necessary information except (as usual) the actual clock speeds this card ships with.

​

As is usual for most MSI cards, the internal packaging for this one is extremely well done with high density foam padding on every side and the accessories pushed out to the side.

​

The accessories list is basic as is the case with most other GTX 460 cards. You get two power adaptors, a DVI to VGA connector and a DVI to HDMI dongle. Conspicuous by its absence is a mini HDMI to HDMI cable or adaptor like most other companies ship with their cards but we are assuming that MSI expects the included DVI to HDMI dongle to sub in.

There’s also a warranty information card with all of the information you will have to give to MSI’s customer support in case you ever have to RMA the card.

 

[SKYMTL]

A Closer Look at the MSI GTX 460 Cyclone OC

A Closer Look at the MSI GTX 460 Cyclone OC

​

MSI has decided to keep the reference design of the 8 ¼” long GTX 460 even though they make mention of their “Military Grade Components” in the marketing materials. As you will see in a shot down below, the components used for the power distribution phases are a bit beefed up.

​

​

Naturally, Cyclone heatsink absolutely dominates all of the visual aspects of this card. It uses a massive 90mm fan mounted amid a radial fin array which blows air in (literally) every possible direction. Due to this omni directional airflow pattern, the GDDR5 memory modules don’t need any of their own heatsinks as they are actively cooled by the fan. To us this is one stunning card.

​

Both the 3+1 VRM design and the components used are the same as the reference design but the actual VRMs are slightly higher spec while the capacitors are Nippon-Chemi Cons while the original card used Lelons. Nonetheless, the Cyclone maintains the status quo with a pair of 6-pin PCI-E connectors.

​

The backplate looks identical to the reference design other than the inclusion of a stylized vent grille. There is a pair of DVI connectors as well as a single mini HDMI output.

 

[SKYMTL]

Test System & Setup

Test System & Setup

Processor: Intel Core i7 920(ES) @ 4.0Ghz (Turbo Mode Enabled)
Memory: Corsair 3x2GB Dominator DDR3 1600Mhz
Motherboard: Gigabyte EX58-UD5
Cooling: CoolIT Boreas mTEC + Scythe Fan Controller (Off for Power Consuption tests)
Disk Drive: Pioneer DVD Writer
Hard Drive: Western Digital Caviar Black 640GB
Power Supply: Corsair HX1000W
Monitor: Samsung 305T 30” widescreen LCD
OS: Windows 7 Ultimate N x64 SP1


Graphics Cards:

MSI GTX 460 768MB Cyclone OC
EVGA GTX 460 768MB Superclocked
EVGA GTX 460 1GB Superclocked EE
NVIDIA GTX 460 1GB (Reference)
GIGABYTE GTX 460 768MB (Stock)
ASUS GTX 465 1GB (Reference)
NVIDIA GTX 470 (Reference)
Sapphire HD 5850 1GB
XFX HD 5830 1GB


Drivers:

NVIDIA 258.96
ATI 10.6 WHQL


Applications Used:

Aliens Versus Predator
Battlefield: Bad Company 2
DiRT 2
Far Cry 2
Just Cause 2
Metro 2033
Unigine: Heaven


*Notes:

FOR MORE INFORMATION ABOUT OUR BENCHMARKING PROCESS PLEASE SEE THIS ARTICLE

- All games tested have been patched to their latest version

- The OS has had all the latest hotfixes and updates installed

- All scores you see are the averages after 3 benchmark runs

All game-specific methodologies are explained above the graphs for each game

All IQ settings were adjusted in-game

 

[SKYMTL]

Aliens Versus Predator (DX11)

Aliens Versus Predator (DX11)

When benchmarking Aliens Versus Predator, we played through the whole game in order to find a section which represents a “worst case” scenario. We finally decided to include “The Refinery” level which includes a large open space and several visual features that really tax a GPU. For this run-through, we start from within the first tunnel, make our way over the bridge on the right (blowing up several propane tanks in the process), head back over the bridge and finally climb the tower until the first run-in with an Alien. In total, the time spent is about four minutes per run. Framerates are recorded with FRAPS.


1680 x 1050

1920 x 1200

2560 x 1600

 

[SKYMTL]

BattleField: Bad Company 2 (DX11)

BattleField: Bad Company 2 (DX11)

To benchmark BF: BC2 we used a five minute stretch of gameplay starting from the second checkpoint (after the helicopter takes off) of the second single player mission up until your battle with the tank commences. Framerates are recorded with FRAPS.


1680 x 1050

1920 x 1200

2560 x 1600

 

[SKYMTL]

DiRT 2 (DX11)

DiRT 2 (DX11)

Being one of the newest games on the market, DiRT 2 cuts an imposing figure in terms of image quality and effects fidelity. We find that to benchmark this game the in-game tool is by far the best option. However, due to small variances from one race to another, three benchmark runs are done instead of the normal two. It should also be mentioned that the demo version of the game was NOT used since after careful testing, the performance of the demo is not representative of the final product. DX11 was forced through the game’s config file. In addition, you will see that these scores do not line up with our older benchmarks at all. This is due to the fact that a patch was recently rolled out for the game which included performance optimizations in addition to new graphics options.

1680 x 1050

1920 x 1200

2560 x 1600

 

[SKYMTL]

Far Cry 2 (DX10)

Far Cry 2 (DX10)

Even though Far Cry 2 has its own built-in benchmarking tool with some flythroughs and “action scenes”, we decided to record our own timedemo consisting of about 5 minutes of game time. It involves everything from run-and-gun fights to fire effects. The built-in benchmarking too was then set up to replay the timedemo and record framerates


1680 x 1050

1920 x 1200

2560 x 1600