DV Hardware - bringing you the hottest news about processors, graphics cards, Intel, AMD, NVIDIA, hardware and technology!
   Home | News submit | News Archives | Reviews | Articles | Howto's | Advertise
DarkVision Hardware - Daily tech news
February 19, 2020 
Main Menu
News archives

Who's Online
There are currently 213 people online.


Latest Reviews
Ewin Racing Flash gaming chair
Arctic BioniX F120 and F140 fans
Jaybird Freedom 2 wireless sport headphones
Ewin Racing Champion gaming chair
Zowie P-TF Rough mousepad
Zowie FK mouse
BitFenix Ronin case
Ozone Rage ST headset

Follow us

NVIDIA GM107 Maxwell architecture revealed

Posted on Thursday, February 13 2014 @ 14:27:12 CET by

VideoCardz claims most of the previous rumors about the GM107, NVIDIA's first Maxwell GPU, were wrong. The site says the core count read by GPU-Z was wrong because the architecture has been changed compared to Kepler. GPU-Z assumed each SMX holds 192 CUDA cores but the actual number is 128. This means the GTX 750 Ti (the full GM107) has 5 SMs with 128 CUDA cores, giving a total of 640, whereas the GTX 750 has 512 (4 SMX).

The GM107 has a TDP of just 60W with most manufacturers will still add a PCIe power connector, for the sake of stability or increasing overclocking headroom. To give you an idea about the power efficiency of this card, VideoCardz notes that the GM107 offer similar performance to the GF100 "Fermi" (GTX 480) GPU, while consuming 4x less power. Maxwell is said to deliver twice the performance per Watt of Kepler!

The GM107 has a die size of 148mm², roughly 30 percent smaller than the GK107 but with a 15 percent higher transistor density. Further details can be read over here.
What is new in Maxwell

Okay, so the SM has been redesigned, the layout has changed, but what exactly makes the Maxwell better? Well here’s the list:

  • Larger L2 cache. This is the main difference between Kepler and Maxwell. Larger L2 cache will limit the queries to the GPU. GM107 L2 cache has 2MB. GK107?s cache has 256KB.
  • Workload balancing and compiler-based scheduling has been improved.
  • The number of instructions per clock cycle has been increased.
  • SM has been redesigned into four processing blocks (as explained above).
  • Maxwell introduces even faster H.264 encoding and decoding with improved NVENC (which is used, for instance, in ShadowPlay).
  • New GC5 power state (low sleep state).
  • NVIDIA GM107 layout



    DV Hardware - Privacy statement
    All logos and trademarks are property of their respective owner.
    The comments are property of their posters, all the rest © 2002-2019 DM Media Group bvba