Recent content by Loki726

L
If Fury is overpriced, why is a $1000 Titan-X acceptable?

There is a reasonably sized market for GPUs with fast single precision and high capacity (e.g. 12GB) memory outside of gaming. Google, Baidu, Facebook, Yandex, Amazon, and others buy them for CUDA and machine learning. At GTC, the NVIDIA CEO said that Titan was targeted at this market...
- Loki726
- Post #143
- Jun 26, 2015
- Forum: Graphics Cards
L
How does a person get an in depth understanding of computers etc??

I mixed self study with formal education. For self study I read as much as possible, of the most detailed material possible. I used the information that I learned to build increasingly complex projects. These in turn exposed me to new problems that required more research to discover...
- Loki726
- Post #11
- Aug 19, 2014
- Forum: Ask a Technical Professional
L
Nvidia Denver... finally here... and it looks good

There was a big difference between Transmeta and Itanium. The Transmeta approach recompiles and reoptimizes the program while it is being executed. The Itanium approach requires the program to be compiled statically. This isn't the same approach at all.
- Loki726
- Post #51
- Aug 14, 2014
- Forum: CPUs and Overclocking
L
NVIDIA preparing four Maxwell GM204 SKUs (VideocardZ via S/A)

I disagree. There is actually a lot of potential for big performance improvements just through better architecture, better circuits, and better software. Think of it this way. Much of architecture, circuits, compiler, and system software research for the past 20-30 years has concentrated...
- Loki726
- Post #259
- Jul 20, 2014
- Forum: Graphics Cards
L
What Is The Minimum Amount of Memory a GPU Should Have?

"640 kB ought to be enough for anyone" comes to mind here... Quite seriously though, it is always possible for someone to write an application/game that does something useful with more memory. So as long as technology scaling continues, there will never be an amount that is high enough for...
- Loki726
- Post #20
- Jul 18, 2014
- Forum: Graphics Cards
L
Single Precision Floating Point in GPUs?

As mentioned above, a single precision floating point value is defined by the IEEE 754 as a 32-bit value that can be manipulated with a specific set of operations, such as multiplication, addition, etc. In GPU specifications, single precision floating point operations per second is commonly...
- Loki726
- Post #8
- Jul 17, 2014
- Forum: Graphics Cards
L
Looking for a special CPU benchmark

Why not just use a game that you think has this behaviour? It is often very difficult to find a benchmark that exactly matches the characteristics of a specific application because there are many factors at play, and the specific factor that limits the performance of your application is not...
- Loki726
- Post #2
- Oct 4, 2013
- Forum: CPUs and Overclocking
L
Linus Torvalds: "Geekbench is SH*T"

I'm constantly surprised by how much we are still learning about a field that has so much history. The application of perfect elimination orders to graph coloring was discovered as a theoretical result in 2003 (http://www.cis.uoguelph.ca/~sawada/papers/chordal.pdf), and the implications of this...
- Loki726
- Post #30
- Sep 27, 2013
- Forum: CPUs and Overclocking
L
Intel TSX Presentation at SuperComputing 2013

Here's a link to the paper: http://pcl.intel-research.net/publications/SC13-TSX.pdf
- Loki726
- Post #14
- Aug 30, 2013
- Forum: CPUs and Overclocking
L
When will we see a completely new Intel uarch (similar to Netburst->Core jump)?

I don't want to derail the thread to go into too much detail on caches However, adding additional levels still adds miss latency since you have to go through another state machine and buffers before hitting the memory controller. Adding an additional layer probably (depending on layout) also...
- Loki726
- Post #48
- Jun 17, 2013
- Forum: CPUs and Overclocking
L
When will we see a completely new Intel uarch (similar to Netburst->Core jump)?

Yeah, I more or less agree that it is possible to increase perf if you are willing to arbitrarily increase cost. We probably will see small 3-5% improvements over time. There may even be a few ideas that can deliver more than this. This is what I meant when I said that there isn't a lot of...
- Loki726
- Post #45
- Jun 17, 2013
- Forum: CPUs and Overclocking
L
When will we see a completely new Intel uarch (similar to Netburst->Core jump)?

I meant within the design constraints of a desktop CPU that can be sold at a profit. If you throw cost out the window you could probably do somewhat better, although this isn't always true since bigger structures often have higher latency. You can never throw out power as a constraint in a CPU...
- Loki726
- Post #43
- Jun 17, 2013
- Forum: CPUs and Overclocking
L
When will we see a completely new Intel uarch (similar to Netburst->Core jump)?

JimmyG is exactly right. There is not a lot of headroom left to increase single threaded CPU performance beyond Haswell barring a major breakthrough (and btw, most architecture research labs have already decided that this problem isn't solvable and have moved onto other problems like power...
- Loki726
- Post #41
- Jun 17, 2013
- Forum: CPUs and Overclocking
L
Microwave Circuit Design vs. Traditional Analog

No, frequency used in the context of microwave means the frequency of the electromagnetic waves. Recall that for EM waves, frequency directly corresponds to wavelength because of the fixed speed of light, so the name refers to the wavelength -- seemingly micro as in micrometer wavelength(10^-6...
- Loki726
- Post #11
- Dec 11, 2008
- Forum: Ask a Technical Professional
L
Power PC

Glad to help if I can. Just try to keep in mind when you are reading posts like this that everyone who participates in these debates (me included) has some agenda that they are trying to push to other people. Even with an in-depth understanding and complete performance characterization of...
- Loki726
- Post #19
- Jan 17, 2007
- Forum: Ask a Technical Professional

RESOURCES

Top Bottom