I must admit that IDF has its moments. I could never imagine that I, a die-hard Apple II fanatic, would ever have the opportunity to meet Steve Wozniak and get an authographed copy of the Apple 1 schematics!
The opening keynote of this year’s IDF lead by Craig Barrett was really disappointing. Usually the opening keynote summarizes new technologies Intel is developing and what we should expect to see in more details during IDF, but this time Barrett spent his time mostly criticizing US educational and health care systems – I agree 100% with Barrett’s remarks, but honestly I was there to hear about new technology and not to learn that Barrett and I agree on our point of view for American society. He also presented some initiatives like microcredit for underdeveloped countries in Africa that kept me asking myself if I were at the right place.
Only one demo was done – in past IDFs tons of demos are done during the keynote presentation – and this was somewhat interesting. A young researcher from Carnegie Mellon University, Dr. Johny Chung Lee, created a USD 50 wireless blackboard by hacking a Nintendo Wii. Adding an inexpensive infrared LED (plus a resistor, switch and battery) to a traditional white board marker, you can make Wii to follow the source of infrared light, thus moving the cursor on the screen according to the movement of the electronic marker. You can see this project (with videos) in details here, including how to make one yourself.
click to enlarge Figure 1: Johny Chung Lee and Craig Barrett demonstrating the wireless blackboard.
In summary, this year’s keynote presentation was a far cry of what it used to be, and that is the reason why I am posting my personal comments about it in form of a blog entry instead of a full article like in the past.
During these three days of IDF I hope to learn more about the architecture of the forthcoming Nehalem CPU (now called Core i7), the architecture of Atom CPU and Intel’s first media CPU for consumer electronics products. Stay tuned.
In order to make the navigation on my blog easier, I've just added navigation buttons on the bottom of the page, similar to the ones we use on our articles. This will help people interested in reading older posts. Enjoy!
As many of you already know, I train Taekwondo since 1992, being a 2nd degree black belt. Last weekend I went to the World Taekwondo Hanmadang 2008, held in Anaheim, CA.
Hanmadang is a different kind of competition, focusing on the martial art aspect of Taekwondo (Mudo, in Korean), like forms (poom-sae) and breaking (kiok-pa), which are the aspects of Taekwondo that I like the best.
I was very excited in competing, as this would be my first international-class competition.
I was really disappointed with the organization. On the first day nobody knew when exactly I was going to compete (which day). My name was in the wrong division for the fist breaking and my name wasn’t on the list for the finals on this category, even though I was a finalist. Then after getting 3rd place, my name wasn’t on the list for getting the medal! For the poom-sae competition I had to wait from 1:00 PM to 9:30 PM. Unbeliveable. Basically they decided to put kids and adults to compete at the same day, so we had to wait until all kids were finished. They could simply put the kids to compete on a different day. On the good side, the location was perfect and many could arrange a vacation to Disneyland while in Anaheim for the competition (Disneyland’s main gate is within walking distance from the convention center).
I always try to see the bright side of everything. I decided to see this tournament as a hands-on training experience for the next one – because I know that I could do a lot better, see below – and also training for my 3rd degree test, which I should take in Q2 2009.
Below I describe my experience on all categories that I participated, all on the senior (31-40 years old) division:
Knife-hand breaking: I didn’t have a clue on the material they would be using. On the official website was written that the material would be granite and I didn’t know exactly what kind of granite they would be using. Of course they chose the hardest granite, the same kind used on kitchen countertops, 1-inch thick. I couldn’t break even one. Now that I know the material I will train for next year or will give up this category. The person that broke the most granite slabs was a Korean master, which was able to break 10 of them. Simply amazing. All competitors gave a standing ovation to him.
Fist breaking: On the preliminaries I was able to break four out of seven plastic tiles, which allowed me to move to the finals. On the finals I broke eight out of ten “real” tiles. I got third place (bronze medal) with this performance. The first place was able to break ten tiles and the second place, nine. Below you can see the videos for both the preliminaries and the finals. I think I could break all ten tiles, I know what my mistake was and I should correct it for next time (in fact watching my performance on the preliminaries I could clearly see that my left foot was on the wrong place and because of that my fist slide when it hit the tiles; I fixed this on the finals).
Foot breaking: On the preliminaries I was able to break four plastic boards, which allowed me to move to the finals. Plastic boards are harder to break than wooden ones. In fact, 2/3 of the competitors couldn’t break the boards on the preliminaries. On the finals I made the mistake of putting too many wooden boards to break (ten), and I wasn’t able to break any. The first place broke nine wooden boards. I think that I can improve a lot here. I was overconfident because I broke the hard plastic boards on the preliminaries and that was my biggest mistake. On the video below you can see my performance on the preliminaries.
Poom-sae: I performed Keumgang, which is the form for the 2nd degree black belt. My biggest mistake was shaking a lot on my last hakdari seogi (crane stance); even if you are a layman you can clearly see that I am shaking a lot at the end of the form. Another thing that probably cost me some points was that instead of answering “yes, sir!” when the referee asked me if I was the competitor shown on the screen I made a thumbs up signal (which can be clearly seen at the beginning of the video below). I got an 82.00 score and a 4th place. Watching the video I can see a lot of other minor mistakes; I could do a lot better. I will train a lot for my next poom-sae tournament
People say that one image is worth more than one thousand words. I think it is the case with this picture of what is below my main bench.
click to enlarge
Plus what is inside my closet.
Yep, I have around 30 hardware parts for reviewing here... This means it would take 1 month reviewing one part per day and asking the manufacturers to stop sending me stuff. Who said running a reviewing website was an easy job?
I really don’t know about you, but sometimes I get the impression that nVidia likes to launch one new video card per week. Even for us sometimes it is hard to keep track of all new releases. The new GTX 200 was launched only three days ago and today nVidia launched a “new” card, GeForce 9800 GTX+ (an overclocked 9800 GTX), lowering the price of GeForce 9800 GTX to make it a competitor to the new Radeon HD 4850. Fortunately nVidia provided a handy table comparing which products from AMD competes to which products from nVidia after this price drop and the addition of GTX+, see below.
click to enlarge
But today nVidia also made an important announcement, which is finally the support of PhysX in the GPU. PhysX is a physics engine, which allows games to be more realistic by calculating how objects interact. For example, if you shot an object, what will happen to the object? Will the object move? Will the bullet make a hole on the object? Will the bullet bounce back? Will the object be destroyed? These calculations are traditionally done by the system CPU. The new 177.39 driver allows GeForce 9800 GTX, GTX 280 and GTX 260 to move these calculations to the GPU, thus improving performance. nVidia is promising for July a new driver that will allow PhysX for the entire nVidia product line.
There are several games supporting PhysX on the market, Unreal Tournment 3 is one of them. 3DMark Vantage also supports PhysX, but it isn’t a real game. Forthcoming games supporting PhysX includes Mstar, Mirrors Edge, Empire: Total War, Backbreaker and Pwnage.
The only problem for me with this amount of new releases in such a short period of time is trying to put together reviews, especially when nVidia launches a new driver version a day, forcing us to retest all video cards everytime a new major driver is released. For instance, with this new driver supporting PhysX released, all our data we had already collected for 9800 GTX and GTX 200 is worthless and even our GTX 280 review published only 3 days ago may be considered "flawed" by criterious readers because of a new driver released today supporting PhysX. This really sucks.
Deciding what methodology to use when reviewing hardware parts like video cards, processors and motherboards is always hard. No matter what methodology we decide to use there will always be users complaining that we should used this or that hardware part, that we should have included this or that program or game, and so on.
There are several technical reasons involved on how we pick the components and programs to do our benchmarking; this selection is not random and also isn’t based on “we used these components because these are the components we have here at our lab”. In fact every time we decide to update our testing procedures we spend a lot of money buying new parts and programs.
Since we are going to start reviewing high-end video cards again – finally! – we decided to explain all the reasons behind the components we picked.
The main goal with a VGA benchmarking is to test the maximum performance it can deliver. If we install a high-end video card on a low-end or mainstream system, the system may limit the maximum performance the video card can deliver because the CPU, the memories and/or the hard disk drive may be limiting the performance (this happens because if the video card is faster than the other components it will have to wait for data that is coming from outside, thus reducing its performance; we want a system where data is available as soon as the video card requires them).
Also since high-end video cards are expensive, we believe a user that buys a high-end card will install it on a high-end PC. Even though there are probably users that install high-end video cards on low-end PCs, it doesn’t make sense to review a high-end video card on a low-end system because the system would be limiting the video card performance, as explained.
So we tried to build the most high-end system we could, i.e. the fastest CPU, the fastest memories and the fastest hard disk drive we could buy. We know that our system isn’t a typical PC, but at least we know for sure that no part from our system is limiting the maximum performance the video card can deliver.
Below we are listing all components from the hardware setup we will use for reviewing high-end video cards, with a brief explanation on why we picked each component and how much each one costs.
CPU: Core 2 Extreme QX9770 (3.2 GHz, 1,600 MHz FSB, 12 MB L2 memory cache). As explained we wanted the fastest CPU around to make sure that the CPU wouldn’t be limiting the video card’s performance. This baby costs USD 1,470.
Memories: Crucial Ballistix PC3-16000 2 GB kit (BL2KIT12864BE2009). We wanted the fastest memories around and these DDR3 memories from Crucial run at 2,000 MHz with 9-9-9-28 timings are are compatible with nVidia’s EPP standard. The biggest problem is the price, USD 500, which prevented us from having a system with 4 GB instead of “only” 2 GB.
Motherboard: EVGA nForce 790i Ultra SLI. We picked this motherboard because it is compatible with our CPU, it is based on PCI Express 2.0 and it can make our memories to run at 2,000 MHz with no fancy configuration. It costs USD 350.
Video monitor: Samsung SyncMaster 305T (LCD, 30”, 2560x1600). We wanted a video monitor that could reach a resolution of 2560x1600 and this was our best choice. We could have bought a cheaper CRT solution, but we wanted a monitor that wouldn’t take a lot of space on your bench. It costs USD 1,230. Notice that we didn’t buy this monitor because of its screen size (30”) but because of its maximum resolution. Notice that there are bigger monitors around that are cheaper, but they have a lower resolution.
Power supply: OCZ EliteXStream 1,000 W. We know that 1,000 W is overkill, but we wanted to make sure that we wouldn’t have any kind of power shortage. It costs USD 250.
So we are talking about a USD 4,100 system. We are glad that except for the video monitor and for the hard disk drive all other parts we got from the manufacturers that are proudly sponsoring Hardware Secrets. Even with these donations we still had to invest USD 1,530 (plus shipping) for you to get the best possible high-end VGA reviews.
Last week during nVidia Editor’s Day Spring 2008 Ubisoft made a gameplay presentation of Far Cry 2. I happened to be there and record the whole presentation, basically the folks from Ubisoft playing the game and showing what’s new – you can watch it below. Funny enough they said that this was the first public presentation of this game, but we could find a few other videos with gameplay presentations of Far Cry 2 on YouTube (but since these other videos were all shot in 2007 we believe our demo was performed with a more updated version of the game). They also provided to the folks present a fact sheet, two high-res screenshots and a technical video presentation, all of which I also included here. Enjoy!
Obs: since YouTube has now a 10-minute limit per video I had to split the video into two parts.
Far Cry 2 gameplay Demo, presented during nVidia Editor's Day Spring 2008 (May 22nd, 2008)
Far Cry 2 technical Demo, showing the new technical features used in the game
Last week nVidia held their Spring 2008 Editor’s day, where they presented their forthcoming series of graphics processing units, which will be released next month. While we can’t talk about this new chip series yet due to the Non-Disclosure Agreement (NDA), we can talk about some ideas that nVidia is seeing as “the future of computing” – basically more GPGPU usage (i.e. the use of the graphics chip to process regular programs) and the co-existence of “competing” technologies like ray tracing and rasterization.
click to enlarge Figure 1: The future of computing, according to nVidia.
During the whole Editor’s Day nVidia repeated ad nauseum how marvelous GPGPU is, showing several examples of applications where performance increased monstrously by the use of this technique. For those who don’t know the concept, the idea is to make the video card GPU to process regular programs instead of using the CPU. What is allowing this to be possible is nVidia’s CUDA compiler, which is capable of compiling any program written in C to be run on any nVidia GPU from series 8 on. We’ve already wrote an article explaining more about this technology.
Is nVidia saying that in the future GPUs will replace CPUs? No exactly. The computer will still need a CPU, but the way nVidia is seeing it the role of the CPU will dramatically decrease in the future. In fact, this is already happening. Thru their “nTeresting” newsletters nVidia has been hammering Intel in the past month, claiming that contrary to what Intel wants you to believe CPU’s aren’t playing an important role on gaming performance anymore and the savvy user should buy a cheaper CPU and spend the saved money on a better video card for a better gaming performance. With GPU performance increasing and more and more tasks that were previously performed on the CPU being transferred to the GPU, this idea makes sense.
With GPGPU this idea will also be valid for regular applications, as soon as mainstream products start to use the GPU for processing, boosting the application performance. During the event Adobe declared that they will start using GPGPU on their forthcoming products, in particular the next version of Photoshop to be released around September, and this could represent the first step towards that direction. Even though from the presentations it is clear that GPGPU can really boost performance on specific applications, the future is always unclear and the performance increase brought by GPGPU for the average user will only depend on software developers upgrading their programs to support it.
It is always important to remind that programs compiled to use GPU’s from nVidia for processing won’t run on ATI’s. Of course final products can detect which video card you have installed on your system and load the CUDA-compiled code – which will provide the performance increase – if you have an nVidia video card.
As for the rasterization vs. ray tracing battle, nVidia is seeing the co-existence of both technologies in the future, as ray tracing is in fact a better technology for some applications, but worse for others. They used as an example the film industry, which use both technologies on movies depending on what needs to be rendered. Nowadays ray tracing isn’t used by games. So by this talk we may expect future GPUs and games to support ray tracing? Maybe, but this probably won’t happen before 2010.
