The Radeon HD 5830

The Radeon HD 5830

Right, let's start off with an overview of the Radeon HD 5830 as a product in general. In our previous articles on almost every Radeon series 5000 card we gave you a pretty deep insight into what the Radeon HD 5800 series is all about. What you guys need to remember is that the Radeon HD 5830 is 100% similar to the feature set of the Radeon HD 5850 and 5870. As such we'll be going a little less deep in this article, yet focus more on specifications, performance, layout and board design.

To understand the Radeon HD 5830, 5850 and 5870 we need to explain a thing or two though. The new Radeon series 5000 products are tagged under the family name 'Evergreen' and within that Evergreen family over the past few months we've seen several products released. The product as tested today falls under the ASIC codename Cypress LE. The Cypress series is the single-chip high-performance graphics solution based on the RV870 chip empowering the Radeon HD 5830, 5850 and 5870.

The Radeon HD 5870 went on sale for $379, while the cheaper entry in the 5800 series, the Radeon HD 5850 is priced at $259. Both cards are fast, really fast. To fill a gap in-between the 159 USD Radeon HD 5770 and 259 USD Radeon HD 5850 AMD released the 5830 priced (initially) at 249 USD.

So now we have established product positioning and price points, let's have a peek at some of the key features for this product series:

• 1GB GDDR5 memory
• ATI Eyefinity technology with support for up to six displays (three on default reference cards)
• ATI Stream technology
• Designed for DirectCompute 5.0 and OpenCL
• Accelerated Video Transcoding (AVT)
• Compliant with DirectX 11 and earlier revisions
• Supports OpenGL 3.1
• ATI CrossFireX multi-GPU support for highly scalable performance
• ATI Avivo HD video and display technology
• Dynamic power management with ATI PowerPlay technology
• DL-DVI, DisplayPort, HDMI
• PCI Express® 2.0 support

We'll address the preponderance of features in our article, but let's focus on the sheer technical specifications, transistor count for example.

The number of transistors on the 5830 is not at all different from the 5870/5870. All products have the same GPU, yet the 5830 has three out of the ten available clusters of shader processors disabled.

We went from 800 Shader processors on the Radeon HD 4850/4870/4890 to 1600 shader processors (also called stream processors) on the fastest single GPU based graphics card, the Radeon HD 5870. And even this much cheaper 5830, still has 1120 Shader processors under its hood.

• Radeon HD 5830  has 7x 160 Shader clusters = 1120 Shader processors
• Radeon HD 5850  has 9x 160 Shader clusters = 1440 Shader processors
• Radeon HD 5870  has 10x 160 Shader clusters = 1600 Shader processors

Why are these clusters disabled? Well, they very well could be wafer yields with an error. Graphics architecture is redundant with failovers. Much like a truck with, in this case, 10 wheels; if two tires are flat, the driver can raise the wheels, not having them touch the road, rendering them inoperable and the truck still can move on, though slower. In AMD ATI's case they simply disable the affected clusters in that processor die.

What happens more often though is that the clusters (while working fine) are simply disabled, just to create enough volume and satisfy demand in the market. If at one point your yields go up and you have plenty of stock, well you do want to sell such parts instead of have them lying around.

These disabled clusters are disabled by laser cuts btw, it's fairly safe to say there's not a chance, like in the good old days of softmodding them or flashing a Radeon 5870 BIOS to the card to get a full 1600 shader processors available (you never know for sure though).

So while the Radeon HD 4870 had 956 million transistors embedded onto its die, the new Radeon HD 5800 GPUs have 2.15 billion transistors. Yes Sir, that is 2150 million transistors tucked away in a small chip. The fabrication node again is 40nm for this product, resulting in a die size of 334 mm².

Now you'd think with so many transistors high clock frequencies would be an issue. Incorrect, the high-end Radeon HD 5830 will be clocked at a steady 800 MHz on its core and shader processor domain. The GDDR5 memory is clocked at 1000 MHz (4000 MHz effective). Both can be overclocked higher though... more on that later, of course.

But before you get blinded by all the specs in a few lines of text, let's break down the two new cards announced in comparison to previous generation Radeon HD 4870.

As you have been able to observe, the 5830 has three shader processor partitions less than the 5870. We really expected this to have been two resulting in 1280 Shader processors.

Also ATI knocked off the number of ROPs (Render Output Units also known as Raster Operation Pipelines) as well, now half at only 16 units. As a result of these factors the pixel fillrate is cut down tremendously, more than we expected for this product really.

To compensate, the 5830 clock frequency has been increased to 800 MHz, where a 5850 runs at 725 MHz. Another nasty side effect of the higher clock frequency is an increase in power consumption. The 5830 consumes more power than a 5850 (!) while performing slower. Very weird.

MSI's R5830 Twin Frozr II edition

So we spoiled the beans already in the introduction. But with the Twin Frozr II edition MSI certainly tries hard to do something special with the graphics card. You'll see a custom PCB, using only the best quality components like Hi-C Caps to provide more precise GPU voltage, SSC (Solid State Chock) No more Buzz noises coming from your graphics card and of course all Solid capacitors. But any graphics cards these days used them standard.

Let's have a look at some of the specifications:

Twin Frozr II Thermal Design

• Dual 8CM PWM Fans
• Quad Heatpipes for better heat dissipation
• 12c cooler compared to reference cooling 

Military Class Components

• Hi-C Cap provides more precise GPU voltage and has better stability.
• SSC (Solid State Chock) for No Buzz noise.
• All Solid CAP for longer lifespan

Over voltage function of GPU

• Use afterburner to adjust GPUs voltage, voltage is up to 1.35V.

Eyefinity + Stream + DirectX 11

• ATI Eyefinity supports 3 monitor outputs simultaneously. (one of them must be DisplayPort)
• Support the latest DX11 games for best gaming experience
• ATI Stream technology can use GPUs ability to co-work with CPU to improve performance.

Overall, the card just looks amazing and seems to be built out of the very best components. But are there any changes compared to the reference model ?

As you can see, there literally will not be a performance difference in-between the reference product and this one. A bit of a shame really as the potential with this card is just so much higher, but ATI does not allow increased clock frequencies. Once you start to overclock you'll notice that there is plenty of performance left in MSI product alright, but of course we'll show you that later on in the review as well.

Let's head on over to the product gallery and look at the product in way more detail guided by a lot of photo's.