XFX Radeon HD 5750 Video Card Review
By Gabriel Torres on October 21, 2009
Radeon HD 5700 series is the first mid-range DirectX 11 video cards to arrive on the market. Let’s see the performance of Radeon HD 5750 from XFX and see if it is a good buy.
Radeon HD 5750 runs internally at 700 MHz and features 720 processors – a little bit less than the 800 processors found on Radeon HD 4850 –, accessing its memory through a 128-bit interface, which is half the width found on Radeon HD 4850. Radeon HD 5750, like Radeon HD 4870 and Radeon HD 5870, uses GDDR5 memory chips, which are capable of transferring four data per clock cycle instead of just two. This makes the memories, which are accessed at 1,150 MHz, to achieve a performance as if they were accessed at 4.6 GHz, providing a 73.6 GB/s maximum theoretical transfer rate. To compare these specs to other graphics chips please take a look at our AMD ATI Chips Comparison Table and NVIDIA Chips Comparison Table.
The main competitor to Radeon HD 5750 is GeForce GTS 250. In our review we will be mainly comparing the performance from these two video cards. We also included in our review Radeon HD 4850 and Radeon HD 4870, plus the new Radeon HD 5770 and GeForce GTX 260/216, and we should come up with a very interesting comparison.
Now let’s talk specifically about the Radeon HD 5750 from XFX. As you may be aware, XFX – which, by the way, many years ago was known as “Pine” – was for years one of the leading NVIDIA partners, and a while ago they decided they shouldn’t manufacture only NVIDIA-based video cards.
This model, also known as HD-575X-ZNFC, comes with 1 GB, two DVI outputs, one HDMI output and one DisplayPort output, following the reference model from AMD.This video card allows you to use up to three video monitors at the same time as a single desktop, feature known as “Eyefinity.” But there is a catch: the third monitor must use the DisplayPort connector, which is still not popular.
We removed the video card cooler to take a look. As you can see in Figure 5, the cooler is very simple, using an aluminum base. It doesn’t touch the memory chips and they have no heatsinks on them.
With the cooler removed you can see that almost all capacitors are solid, which is a terrific feature as they don’t leak (the ones that are not solid are from Elcon, a Chinese manufacturer). Like the memory chips, the transistors from the voltage regulator don’t have a heatsink.
XFX Radeon HD 5750 uses eight 1 Gbit Hynix H5GQ1H24AFR-T2C GDDR5 chips, making its 1 GB memory (1 Gbit x 8 = 1 GB). These chips have a maximum transfer rate of 5 Gbps (“T2C” marking), which is equivalent of a 5 GHz GDDR5 clock or 1.25 GHz (5 GHz / 4) real clock. Since on this video card the memory was running at 1.15 GHz, there is an 8.7% headroom for you to overclock the memories with them still running inside their specifications. Of course you can always try pushing them above their specs. In Figure 7 we provide a close-up of the GDDR5 memory chips.
In Figure 8, you can see all accessories that come with this video card, basically a quick installation guide, a driver CD, a power adapter for converting two peripheral power plugs into one six-pin video card power connector, and a “do not disturb” sign.
XFX Radeon HD 5750 main features are:
* Researched at Newegg.com on the day we published this review.
During our benchmarking sessions, we used the configuration listed below. Between our benchmarking sessions the only variable was the video card being tested.
We adopted a 3% error margin; thus, differences below 3% cannot be considered relevant. In other words, products with a performance difference below 3% should be considered as having similar performance.
3DMark Vantage measures Shader 4.0 (i.e., DirectX 10) performance and supports PhysX, a programming interface developed by Ageia (now part of NVIDIA) to transfer physics calculations from the system CPU to the video card GPU in order to increase performance. Mechanical physics is the basis for calculations about the interaction of objects. For example, if you shoot, what exactly will happen to the object when the bullet hits it? Will it break? Will it move? Will the bullet bounce back? Note that since we are considering only the GPU score provided by this program, physics calculations are not taken into account.
We ran this program at three 16:10 widescreen resolutions, 1680x1050, 1920x1200, and 2560x1600. First we used the “Performance” profile, and then we used the “Extreme” profile (basically enabling anti-aliasing at 4x, anisotropic filtering at 16x, and putting all detail settings at their maximum or “extreme” values). The results being compared are the “GPU Scores” achieved by each video card.
Radeon HD 4850 and Radeon HD 4870 didn’t produce a reliable score for the “Extreme” profile under 2560x1600, so these particular results should not be considered.
|3DMark Vantage - Performance - 1680x1050||GPU Score||Difference|
|GeForce GTX 260/216||7167||32.06%|
|Radeon HD 4870||7135||31.47%|
|Radeon HD 5770||6831||25.87%|
|Radeon HD 4850||5477||0.92%|
|Radeon HD 5750||5427|
|GeForce GTS 250||5148||5.42%|
|3DMark Vantage - Performance - 1920x1200||GPU Score||Difference|
|GeForce GTX 260/216||6860||61.34%|
|Radeon HD 4870||5598||31.66%|
|Radeon HD 5770||5368||26.25%|
|Radeon HD 5750||4306|
|Radeon HD 4850||4252||0.00%|
|GeForce GTS 250||3893||9.22%|
|3DMark Vantage - Performance - 2560x1600||GPU Score||Difference|
|GeForce GTX 260/216||3840||54.47%|
|Radeon HD 4870||3230||29.93%|
|Radeon HD 5770||3048||22.61%|
|Radeon HD 5750||2486|
|Radeon HD 4850||2380||4.45%|
|GeForce GTS 250||2149||15.68%|
|3DMark Vantage - Extreme - 1680x1050||GPU Score||Difference|
|GeForce GTX 260/216||5615||37.93%|
|Radeon HD 5770||5109||25.50%|
|Radeon HD 4870||5093||25.10%|
|Radeon HD 5750||4071|
|GeForce GTS 250||4000||1.77%|
|Radeon HD 4850||3941||3.30%|
|3DMark Vantage - Extreme - 1920x1200||GPU Score||Difference|
|GeForce GTX 260/216||5240||60.79%|
|Radeon HD 5770||4106||25.99%|
|Radeon HD 4870||4040||23.96%|
|Radeon HD 5750||3259|
|Radeon HD 4850||3119||4.49%|
|GeForce GTS 250||3073||6.05%|
|3DMark Vantage - Extreme - 2560x1600||GPU Score||Difference|
|GeForce GTX 260/216||3020||58.36%|
|Radeon HD 5770||2417||26.74%|
|Radeon HD 5750||1907|
|GeForce GTS 250||1729||10.29%|
|Radeon HD 4870||518||268.15%|
|Radeon HD 4850||402||374.38%|
Call of Duty 4 is a DirectX 9 game implementing high-dynamic range (HDR) and its own physics engine, which is used to calculate how objects interact. For example, if you shoot, what exactly will hapen to the object when the bullet hits it? Will it break? Will it move? Will the bullet bounce back? It gives a more realistic experience to the user.
We ran this program at three 16:10 widescreen resolutions, 1680x1050, 1920x1200, and 2560x1600, maxing out all image quality controls (i.e., everything was put on the maximum values on the Graphics and Texture menus). We used the game internal benchmarking feature, running a demo provided by NVIDIA called “wetwork.” We are putting this demo for downloading here if you want to run your own benchmarks. The game was updated to version 1.6. The results below are the average number of frames per second (FPS) achieved by each card.
|Call of Duty 4 - 1680x1050||FPS||Difference|
|GeForce GTX 260/216||86.44||16.68%|
|GeForce GTS 250||82.34||11.15%|
|Radeon HD 4870||83.80||13.12%|
|Radeon HD 5770||79.86||7.80%|
|Radeon HD 5750||74.08|
|Radeon HD 4850||69.54||6.53%|
|Call of Duty 4 - 1920x1200||FPS||Difference|
|GeForce GTX 260/216||83.50||33.81%|
|GeForce GTS 250||75.42||20.87%|
|Radeon HD 4870||73.60||17.95%|
|Radeon HD 5770||70.14||12.40%|
|Radeon HD 5750||62.40|
|Radeon HD 4850||57.18||9.13%|
|Call of Duty 4 - 2560x1600||FPS||Difference|
|GeForce GTX 260/216||64.36||57.67%|
|GeForce GTS 250||49.60||21.51%|
|Radeon HD 4870||47.40||16.12%|
|Radeon HD 5770||46.08||12.89%|
|Radeon HD 5750||40.82|
|Radeon HD 4850||36.26||12.58%|
Crysis Warhead is a DirectX 10 game based on the same engine as the original Crysis, but optimized (it runs under DirectX 9.0c when installed on Windows XP). We ran this program at three 16:10 widescreen resolutions, 1680x1050, 1920x1200, and 2560x1600, maximizing image quality (16x anti-aliasing, 16x anisotropic filtering) and using the Airfield demo. The results below are the number of frames per second achieved by each video card.
|Crysis Warhead - 1680x1050||FPS||Difference|
|GeForce GTX 260/216||18.0||38.46%|
|Radeon HD 5770||17.0||30.77%|
|GeForce GTS 250||14.0||7.69%|
|Radeon HD 5750||13.0|
|Radeon HD 4850||12.0||8.33%|
|Radeon HD 4870||11.5||13.04%|
|Crysis Warhead - 1920x1200||FPS||Difference|
|GeForce GTX 260/216||15||36.36%|
|Radeon HD 5770||13||18.18%|
|Radeon HD 5750||11|
|GeForce GTS 250||10||10.00%|
|Radeon HD 4870||10||10.00%|
|Radeon HD 4850||8||37.50%|
|Crysis Warhead - 2560x1600||FPS||Difference|
|Radeon HD 5770||8||14.29%|
|Radeon HD 5750||7|
|GeForce GTX 260/216||5||40.00%|
|GeForce GTS 250||4||75.00%|
|Radeon HD 4850||1||600.00%|
|Radeon HD 4870||1||600.00%|
Fallout 3 is based on the same engine used by The Elder Scrolls IV: Oblivion, and it is a DirectX 9.0c (Shader 3.0) game. We configured the game with “ultra” image quality settings, maxing out all image quality settings, at three 16:10 widescreen resolutions, 1680x1050, 1920x1200, and 2560x1600. To measure performance, we used the FRAPS utility running an outdoor scene at God mode, running through enemy fire, triggering post processing effects, and ending with a big explosion in front of Dupont Circle.
|Fallout 3 - 1680x1050||FPS||Difference|
|GeForce GTX 260/216||77.31||8.46%|
|Radeon HD 5770||76.43||7.23%|
|Radeon HD 4870||75.32||5.67%|
|GeForce GTS 250||74.33||4.28%|
|Radeon HD 5750||71.28|
|Radeon HD 4850||69.36||2.77%|
|Fallout 3 - 1920x1200||FPS||Difference|
|GeForce GTX 260/216||77.30||16.91%|
|Radeon HD 4870||73.45||11.09%|
|Radeon HD 5770||70.95||7.30%|
|GeForce GTS 250||68.75||3.98%|
|Radeon HD 5750||66.12|
|Radeon HD 4850||59.56||11.01%|
|Fallout 3 - 2560x1600||FPS||Difference|
|GeForce GTX 260/216||61.49||31.30%|
|Radeon HD 5770||53.91||15.12%|
|Radeon HD 5750||46.83|
|GeForce GTS 250||46.09||1.61%|
|Radeon HD 4870||44.32||5.66%|
|Radeon HD 4850||33.25||40.84%|
Far Cry 2 is based on an entirely new game engine called Dunia, which is DirectX 10 when played under Windows Vista with a DirectX 10-compatible video card. We used the benchmarking utility that comes with this game, setting image quality to the maximum allowed and running the “Ranch Long” demo three times. The results below are expressed in frames per second and are an arithmetic average of the three results collected.
Radeon HD 4850 and Radeon HD 4870 could not run Far Cry 2 at 2500x1600 with all image quality settings maxed out.
|FarCry 2 - 1680x1050||FPS||Difference|
|GeForce GTX 260/216||38.92||58.79%|
|GeForce GTS 250||31.07||26.76%|
|Radeon HD 5770||27.50||12.20%|
|Radeon HD 5750||24.51|
|Radeon HD 4870||11.29||117.09%|
|Radeon HD 4850||9.80||150.10%|
|FarCry 2 - 1920x1200||FPS||Difference|
|GeForce GTX 260/216||33.82||55.35%|
|GeForce GTS 250||26.23||20.49%|
|Radeon HD 5770||24.35||11.85%|
|Radeon HD 5750||21.77|
|Radeon HD 4870||8.94||143.51%|
|Radeon HD 4850||7.38||194.99%|
|FarCry 2 - 2560x1600||FPS||Difference|
|GeForce GTX 260/216||18.61||51.30%|
|GeForce GTS 250||17.15||39.43%|
|Radeon HD 5770||14.49||17.80%|
|Radeon HD 5750||12.30|
Unigine is a 3D engine used by some games and simulations. The developer provides two demos for this engine, Tropics and Sanctuary. We ran the Tropics benchmarking module under DirectX 9 mode at full screen with image quality settings maxed out. The results below are the number of frames per second achieved by each video card.
|Tropics - 1680x1050||FPS||Difference|
|Radeon HD 4870||42.4||24.34%|
|Radeon HD 5770||40.9||30.46%|
|Radeon HD 5750||34.1|
|Radeon HD 4850||32.5||4.92%|
|GeForce GTX 260/216||31.4||8.60%|
|GeForce GTS 250||26.7||27.72%|
|Tropics - 1920x1200||FPS||Difference|
|Radeon HD 5770||34.5||19.79%|
|Radeon HD 5750||28.8|
|GeForce GTX 260/216||28.5||1.05%|
|Radeon HD 4850||27.4||5.11%|
|Radeon HD 4870||24.6||17.07%|
|GeForce GTS 250||21.5||33.95%|
The good news about Radeon HD 5750 is that in the worst-case scenario it achieved the same performance as Radeon HD 4850, being on most situations between 4% and 13% faster than this previous-generation video card. On situations that high-performance is needed, in particular on Far Cry 2 with image quality settings maxed out, Radeon HD 5750 achieved more than double the performance from Radeon HD 4850. And certain configurations that Radeon HD 4850 can’t run (2560x1600 with image quality settings maxed out on 3DMark Vantage and Far Cry 2) the new Radeon HD 5750 can. And the amazing thing is that when Radeon HD 4850 was originally released it cost around USD 200 and the new Radeon HD 5750 costs only USD 140, so you can get similar or higher performance plus DirectX 11 support at a lower cost.
The problem is that on most of our tests the main competitor from NVIDIA, GeForce GTS 250, was faster. Only on 3DMark Vantage (up to 10%) and on Unigine Tropics (between 28% and 34%) Radeon HD 5750 was faster than GeForce GTS 250. For example, on Call of Duty 4 GeForce GTS 250 was between 11% and 22% faster, on Crysis Warhead it was 8% faster and on Far Cry 2 it was between 20% and 39% faster. On Fall Out 3 the difference between the two cards was too small. Keep in mind that on all these games we cranked up all image quality settings to their max.
Radeon HD 5750 is definitely o good replacement for Radeon HD 4850, as explained, but it may fall behind its main competitor, GeForce GTS 250, depending on the resolution and image quality settings you like to play the most. From the six programs we ran, GeForce GTS 250 won on three of them, Radeon HD 5750 won on two of them and they tied on one of them. Since we can’t run all the games available on the market and since our results do not point out a clear winner, we have to claim a technical tie.
We are given our Silver Award instead of our Golden because, as explained, Radeon HD 5750 isn’t faster than its main competitor on all situations.