Should You Buy AMD or NVIDIA?

NVIDIA (NVDA) is a US$ 4,469 million (2015) company and has a total of 8,808 employees working across 54 global offices and located in 24 countries by the end of fiscal year 2014.  The company’s workforce is primarily concentrated in the US/Canada followed by Asia-Pacific, India and Europe/Middle East/Africa.

AMD (AMD) is a US$ 3.991 billion (2015) company and has a total of 9,139 employees working across over 50 global locations and spread through 31 key countries including, North America- US & Canada, Europe, Asia Pacific, South & Central America & Caribbean, Middle East & Africa.

Both the organizations have extremely well-diversified work forces with NVIDIA still a larger company in capital size compared to AMD and has smaller employee base compared to the counterpart, giving it a somewhat greater opportunity to expand the global operations while continuing to minimize non-core spending.

a.) Market Share

NVIDIA’s GPU share in the first quarter of 2016 is still market-leading at over 80% of the worldwide GPU market and well-ahead of AMD’s market share.
However, NVIDIA is still executing on the strategy of providing driver provision for the innovative API in its recently launched cards.

AMD has continued to lose its key market share to NVIDIA despite illustrating higher performance for DirectX 12 enabled gaming titles through an advanced Async ready GCN architecture. During 2015 and on DirectX 11 platform, AMD’s Radeon cards failed to illustrate superior optimizations at the required pace compared to the GeForce cards.

On the drivers’ front, AMD has continued to launch either Alpha or BETA versions of its drivers which have illustrated significantly superior performance with the Radeon cards on the latest AOTS benchmark results. Contrastingly, NVIDIA has uniquely focused on delivering game ready drivers as against Alpha or BETA drivers which might be a reason that they still don’t have new drivers launched in the market.

b.) New Architecture

Going forward, NVIDIA is believed to soon introduce its Pascal GTX GeForce cards with the flagship GPU NVIDIA GP100 and having TSMC 16nm FinFET GPU process with 15.3 billion GPU transistors, based on 5th Gen Pascal CUDA graphics architecture with 1 TB/s of peak HBM2 bandwidth.

Moving ahead, AMD is expected to rock the market with its flagship GPUs Polaris 10 and Vega 10, both having GloFo 14nm FinFET GPU processes. AMD Polaris has TBC GPU transistors based on GCN 4.0 (Polaris) graphics architecture with GDDR5/X of HBM memory at HBM2 bandwidth of 512 GB/s to be launched in 2016 while AMD Vega has 15 to 18 billion GPU transistors based on GCN 4.0 (Vega) graphics architecture with HBM2 memory at 1 TB/s of peak HBM2 bandwidth to be launched in 2017.

Therefore, AMD is estimated to regain its lost market share from NVIDIA with its significantly compact and highly-efficient Polaris GPUs that are equipped with HBM2 memory bandwidth and allowing it to design significant amounts of VRAM on a highly compact chip while consuming notably lesser power compared to the earlier employed GDDR5 chips and thus, enhancing the complete chip’s overall efficiency.

c.) Product Pricing and Costs

Price-wise, both AMD and NVIDIA are somewhat at the same level considering the costs with Nvidia GeForce 980Ti and AMD R9 Fury X costing nearly the same.

Performance-wise, NVIDIA is undoubtedly and consistently the industry-leader with the company’s fastest Titan X graphics card offering boasting of an effective memory speed which is seven times faster than AMD’s latest Fury X graphics card and that too at a comparable price.

Thus, the key gaming enthusiasts must be bullish on NVIDIA, offering superior gaming performance at competitive pricing compared to AMD which is yet to launch the advanced Polaris-based graphics cards to compete effectively with NVIDIA.
Published on Jun 10, 2016
By Yaggyaseni Mittra

Copyrighted 2016. Content published with author's permission.

Posted in ...