Showing posts with label Nvidia. Show all posts
Showing posts with label Nvidia. Show all posts

Aug 7, 2008

Graphics Card Buying Guide


This Buying Guide was published on PC Shopper Philippines magazine (August Issue).

If you intend to use your PC for graphics-intensive applications, the video card is definitely one of the most important components that will make or break your tasks. With the rapid evolution of hardware and software that require constant upgrades, browse through our advices to make sure you will get the right add-on card.

Fast Facts
  • If you’re running 3D and heavy graphics, your system mostly depends on the graphics card. A weak video card will slow down PC performance even if you have a strong CPU or a massive memory module. Other components may complement, but video output quality is loaded based on the graphics card’s ability to handle the technologies. This is why some high-end video cards are actually more expensive than any other components in the system.
  • Graphics cards are usually rolled out between six to twelve months period before a new faster and powerful series than the previous models is introduced to support new technologies. Current mid-range to high-end video cards bring a lot of new highfalutin technologies that are often times blurry for most of us.
  • PC Gaming publishers usually release new products every year, giving way to new system requirements, making your head turning to find out why your PC gets outdated quickly. If your PC is more than three years, maybe it's too old to have an upgrade so it’s better to get rid of it. No, don’t throw it away. Sell it and buy a new rig.
  • Graphics card upgrade may require upgrading of other components such as CPU, memory, hard drive, and even the power supple depending on the applications and video card type you will use.
  • Graphics cards are also called “expansion cards” or “add-on cards” specifically designed for the motherboard’s AGP or PCI-Express slots. They called “expansions” or “add-ons” simply because there are cheaper video solutions that are integrated or built-in on the motherboard with special slots for discreet video cards. Built-in video may run 3D applications and games but this is generally slow and works well on daily office tasks and e-mail access.
AGP and PCI Express
Current motherboards usually have AGP (Accelerated Graphics Port or Advanced Graphics Port) or PCI Express (Peripheral Component Interconnect Express or PCI-E) expansion slots for add-on graphics cards. You should determine the slot type on your motherboard because AGP or PCI Express graphics cards are designed for one slot type only. The big difference between the two is the bandwidth they offer. Manufacturers claim that the higher bandwidth is better.

The descendant of PCI parallel-interface video cards, AGP is being slowly driven away by PCI Express. Providing sufficient bandwidth for new applications with 2.1GBps bandwidth, the latest AGP 8x graphics cards and motherboards are still widely available.

PCI Express is the latest species in the evolution of video cards following on the component transition from parallel to serial for an increased bandwidth. It theoretically offers 2 – 4 times more bandwidth than AGP. Since all PC components are gearing towards the serial interface, most new graphics cards designed for PCIE. The latest PCIE (x16) 2.0 can run about 8 GBps bandwidth. Most motherboards with multiple PCI Express x16 slots are designed to operate SLI and CrossFire functions.

If your PC system is more than two years old, it probably has an AGP slot. Upgrading to PCI Express will be expensive since you'll need to replace the motherboard, CPU, and memory, but if your system is more than two years old, it might just be the right time to upgrade your entire PC anyway.
Graphics Card Stuff: What you may not know
  • Fill Rate – GPU’s speed in drawing pixels depending on the number of ROP (raster operations pipeline) units of the graphics card.

  • GART (Graphics Address Remapping Table) – a capability of reading textures directly from memory without the need to copy them to video memory.

  • GBps (Gygabyte per second) – bandwidth measured in the number of Gygabytes processed per second.

  • Pixel Pipelines –pixels pass through a GPU unit called “pipelines” which crunch complex vertex and pixel-shader programs for lighting and other effects. The more pipelines, the better. Cheaper cards usually have 4 pipelines while mid-range to high-end 8-16 or more.

  • Pixel Shaders (also called “pixel processors”), Vertex Shaders Units (also called “vertex processors”), and Unified Shaders – Usually used for realism gaming, Pixel and Vertex Shaders are GPU units that execute pixel frequency and control colors that apply an effect on images like shadows and explosions. Unified Shaders combines and executes both Pixel and Vertex Shaders.

  • RAMDAC (Random Access Memory Digital-to-Analog Converter) – converters used to store color palette that generate analog signals for the monitor.

  • SLI (NVIDIA’s scalable link interface) and CrossFire (ATI) – New technologies that allow two or graphics cards installation for certain intensive graphics applications.
It’s not all about memory
The most popular error in pronouncing a graphics card’s performance is judging it by its memory size. Well, that was true during PCI video cards heydays. Memory size is important, but the real deal in determining video card speed is the capability of the GPU. Of course, higher memory can provide a certain effect but there are more important considerations to measure performance like the GPU’s memory bus, clock, and pipelines that affect bandwidth. Take a look at pipelines and clock speed specs

Current graphics cards use 64 - 512-bit bus witch 400 – 2000MHz clock, so a 512-bit bus with high clock speed can definitely give the card a boost as compared to 64-bit. Current memory types are designed on DDR and GDDR modules ranging from 128 MB to 1024 MB. A few years ago, video cards with 16-32 MB could do well with graphics programs. Today, many games and applications like Adobe CS3 require at least 128 MB to run.

What’s on the card?

  • Analog D-Sub (also known as VGA or DB-15F) – 15-pin connector for analog monitors.

  • DVI (DVI-Integrated and DVI-Digital) - standard interface used to provide digital video signals to LCD monitors.

  • HDMI (High Definition Multimedia Interface) - simultaneously transmitting visual and audio data via the same cable.

  • DisplayPort - new digital video interface designed for upcoming multimedia devices.

  • GPU (Graphics Processing Unit) – The brain of the graphics card, GPUs operation is measured in megahertz (millions of information processed per second).
nVIDIA or ATi?
ATi and nVIDIA are the only two consumer graphics card vendors that clashes head to head in the market today (If there are others, please let me know). They have their own price ranges and I can’t say which is better or worst because each has its ups and downs. If you don’t have any budget constraints, grab a high-end solution to get the highest quality images and best performance.

Jun 15, 2008

NVIDIA unleashes GeForce GTX 200 Series

Geared up to accelerate heavy graphics tasks while delivering exciting visual experience, NVIDIA recently unleashed its latest GeForce GTX 200 graphics cards series with the company’s new flagship GeForce G200 GPU (graphics processing unit). Manufactured using 65nm process technology featuring an unprecedented amount of transistors for higher performance and enhanced PC gaming performance, the G200 GPU holds 1.4 billion transistors on its 240 unified shader/stream processors (on GTX 280) with re-designed thread dispatch processor, 80 texture units (80 texture address, 80 texture filtering units), 32 render back ends as well as 512-bit memory controller supporting PCI Express 2.0 platform.

The new family of GTX 200 graphics cards include the GTX 280 and GTX 260, all equipped with a new version of the NV IO controller that supports D-Sub, dual-link DVI, DisplayPort, and TV-Out. Other feature highlights of the GTX 280 include 1296MHz processor clock, 1GB GDDR3 standard memory configuration with 1107MHz memory clock, 512-bit memory interface width, and 141.7GB/sec memory bandwidth. The GTX 260 clutches 192 processor cores, 576MHz graphics clock, 1242MHz processor clock, 999MHz memory clock, 896MB GDDR3 standard memory configuration, 448-bit memory interface width, and 111.9GB/sec memory bandwidth.

The GTX 280 and GTX 260 include the second generation NVIDIA Unified Architecture, which delivers more gaming performance over previous NVIDIA GeForce 8800 Ultra through improved 240 processor cores for an enhanced shading horsepower at resolutions as high as 2560 x 1600 dpi.

NVIDIA’s new 3-way SLI and PureVideo HD technologies are built in the GTX 280 and GTX 260 for implementing 3-way alternate frame rendering dubbed “the world’s fastest gaming solution under Windows Vista.” The new GTX 200 series is also set to support the groundbreaking NVIDIA CUDA technology, which now allows programmers to write applications and tinker on the GPU.

Dec 4, 2007

First 1GB on board graphics card launched by Palit

Taiwan-based Palit Microsystems recently introduced what is touted as “the world’s first on board 1GB memory graphics processor.” I was invited in a media briefing on December 3 at MyCinema Greenblet 3, Makati City where the NVIDIA-chipset based Palit GeForce 8500GT Super+1GB was launched to the Philippine market.

According to nVidia executives from Taiwan who was also present during the product launch, the new 8500GT is packed with actually 1024MB king size memory which can boost heavy graphics computing performance and save bandwidth on the memory capacity bottleneck.

The 8500GT is based on PCI Express platform with 128-bit memory interface and a rated 400MHz RAMDACs. RAMDACs (Random Access Memory Digital-to-Analog Converters) is a display adapter’s memory that stores the color palettes to enhance color displays. So the more color a RAMDAC can process, the better.