The AnandTech Podcast: Episode 15

31 Comments

View All Comments

• themossie - Tuesday, January 29, 2013 - link

  UC Davis students and grads (Davis is in Yolo County) are getting a real kick out of the Yolo thing.
  
  Except that we make fun of them for it :-)

• watersb - Tuesday, January 29, 2013 - link

  Very happy with another Anandtech podcast!
  
  I was waiting for you all to connect the dots for us, putting this Intel announcement of no more motherboards together with the previous rumors of Haswell as initially a BGA-only part…

• creed3020 - Tuesday, January 29, 2013 - link

  "Oh look a lion, YOLO!" ~Dr Ian Cutress
  
  Ian that was awesome. Full stop.
  
  I really enjoyed this podcast as the off the cuff comments made the show!

• Krysto - Tuesday, January 29, 2013 - link

  Won't Nvidia become very strong soon in the baseband market with their soft-mdeom Icera 500, which can adapt to any frequency, and basically make phones work on any frequency, without having to implement physical parts for certain bands?

• jeffkibuule - Thursday, January 31, 2013 - link

  SDR can't adapt to any frequency. The software can change the way it processes data from external antennas (like dealing with newer 3GPP standards), it doesn't change the frequency an antenna is tuned for.

• dragon0005 - Tuesday, January 29, 2013 - link

  guys first zolo is not an intel brand coz lava the company which makes zolo phone has another one with an ARM processor
  and the yolo man any time you watch some NGC r discovery or something an they show like tribes it sounds a little like that

• dragon0005 - Tuesday, January 29, 2013 - link

  XOLO

• tipoo - Tuesday, January 29, 2013 - link

  The whole appeal of consoles is to have a unified specification to develop for across a broad install base. If they cut console generations down to 3-4 years, the install base would never get as large as the current generation, which would mean developers would always have a smaller market to target which they obviously do not want. Plus consumers would probably be less than happy to shell over a few hundred dollars ever 3-4 years.
  
  I do think this generation is excessively long, but 3 would be far too short.
  
  And in terms of a midway update to the same console generation, that again is bad for developers as they then have to put the budget into making sure the game performs well on multiple hardware specifications, which reduces some of the appeal of going with consoles (granted two configs is still better than infinity). Same with add on cards. Remember the N64 expansion pack? Remember how they never ever tried that again?

• alwayssts - Wednesday, January 30, 2013 - link

  I imagine what Anand is saying (beyond the PC-standard argument) is a spec needs to be updated before they become irrelevant to the modern ecosystem dictated by shrinking process tech and gpu cadences. I think two full-node shrinks is a fair cadence, assuming a contemporary spec is built around a ~150w discrete graphics card, which seems fair given where it lies in the performance/price/power/yield curve at any given time.
  
  Essentially, if we start with a 3870 (11/2007) it should be updated by around the time we have llano (mid-2011). Any longer and the spec doesn't make sense, and I agree. That may not be 3.5 years, it may be much longer depending on how quickly nodes ramp. How long until 14nm is as prevalent (using tsmc/gf) as 28nm will be at the end of this year?
  
  For instance, if the next consoles use the computational ability equal to a graphics card made to saturate 16 ROPs, say 896sp/1ghz+ (or in console terms probably 1024sp at a lower speed, ie 875mhz+ for yields/power), we should update that spec before AMD/Intel build that spec into a CPU. It's not crazy to believe that will be 14nm for AMD, and a similar time-frame for Intel.
  
  In terms of console add-ons, it would be interesting to see an update that essentially turns it into a 'crossfire' solution. If the new consoles' common spec becomes the standard for 1080p/60fps, would it be so crazy in a few years time to sell an add-on and update the main system as a new sku to allow it to run similar titles more-or-less 4k/30fps (twice the gpu grunt)? Considering how cheap that would be in terms of hardware to provide that upgrade (essentially adding another contemporary low-end gpu to the box to create a mid-range part) and the extreme likelihood AMD will not be changing the GCN structure any time soon, I don't think it's an asinine request.

• scales - Thursday, January 31, 2013 - link

  I agree.
  
  I do not think Anand’s “open console” is a great idea. Anand appears to push for better and better hardware, but he is neglecting half of the situation.
  
  I left most of my computer gaming in favor of a console because I got sick of shelling out 200-300 bucks every 5 months for a new graphics card. The whole point of the console is that the consumer doesn’t have to worry about their hardware not being able to play the game. It is extremely frustrating to sit and fiddle with texture and detail settings just to get a higher FPS. Consumers don’t want to have to keep buying new hardware just to play a newer game.
  
  I think it is also pretty easy to tell that most of these games are NOT optimized for any hardware, just ported across each (look how laggy Skyrim is on the PS3). Additionally, developers need to focus on creating unique and optimized content, rather than release more than 5 iterations of the same first-person-shooter. The beauty of the older “DOS-and-before” era games was that there was far more creativity being put into game development (only a few indie games are doing that now).
  
  On another note, if a console’s life cycle were shortened to 3 years, what happens to all of my previous games that are no longer compatible? Do I have to buy all new controllers too?
  
  Shorter life cycles on consoles just doesn't make that much sense.