The TR Podcast 145: The mailbag and top-tier graphics


Plus an update on SSDs, and Jordan's tech woes
— 9:32 PM on November 14, 2013

The Tech Report Podcast

Date: November 14, 2013
Duration: 1:24:49

Hosted by: Jordan Drake

Co-Hosts: Scott Wasson, Geoff Gasior, and Cyril Kowaliski

Download:
MP3 (61MB)

Subscribe:
RSS (MP3) | RSS (M4A)
iTunes (MP3) | iTunes (M4A)

 

Show notes
We are back (belatedly, we know) with more of the fabulous Tech Report Podcast. This show kicks off in a bit of a sour mood as Jordan shares his latest tech woes and how they've affected podcast production. Then it's on to happier things with a slew of listener tweets and emails. Finally, Scott has two major graphics card reviews, as Nvidia and AMD battle for the crown of "fastest GPU." Before we close the episode, Geoff gives us an update as his SSD endurance project hits the 200TB mark.

Send in listener mail, and we'll answer on the podcast. - jdrake@techreport.com

Follow us on Twitter - Scott - Jordan - Geoff - Cyril - The Tech Report

Listener mail/tweets:

Battlefield 4 vs. Crysis 3 - from Jesse:

"I'm curious to know Scott's opinion on Battlefield 4's graphics versus Crysis 3's. Also, destructibility and physics."

Broadwell? - from @turbobeta:

"Tell us everything you know about Broadwell's CPU. Is this a tiny 5% gain, like Haswell was to Ivy Bridge?"

Hot and loud? - from @T_Money_58:

"How loud and hot is too loud and hot? Also has AMD left any over clocking headroom in 290/290X with stock coolers?"

Broadwell? - from Francis:

"Hey guys! This is Francis, from Brazil, with a simple question for you. My system have a Core i5 3470 running at stock clock and a Geforce GTS450. Two 7200 hard drives and two 12cm fans. Nothing too fancy. It's running pretty stable with a 430w Corsair CX430 v2 power supply. This power supply, according to some brazilian sites reviews is able to output almost 517w and is considered to be really good. Now I'm going to buy a new GPU and I have two options at almost the same price, both Gigabyte Windforce cards: The Geforce GTX660 and the R9 270X. It seems that the 660 uses much less power, so I would be okay with my power supply. I'd prefer instead to buy the 270X, since it seems to be faster, but maybe my power supply would be too weak for it. So the question is - Is this power supply good enough to run the 270X?"

Broadwell? - from Mark:

"Just wanna let u know that Leap chair yall talked about is a very popular chair used by our company(XO communications), we work in telecom support and these exact chairs are standard in the call/repair centers fyi....good chair"

AMD Laptops - from P Jones:

"I wanted to ask why don't we see much more AMD temash/kabini tablets/laptops? I have only read of one (MSI W20 3M) AMD temash tablet device win. With all the Intel bay trail devices, I hope to see comparisons of BayTrail vs. Temash for total system performance. How does the eMMC controller on BayTrail impact system performance? Comparing AMD gpu accelerated performance vs. intel I own the acer v5-122p with a 1.0 Ghz, quad core temash cpu and find that the overall system performance is much better than any clover trail device"

Rethink Frametime? - from Doug:

"I think your frametime based evaluations are the best thing to happen to video card evaluations since sliced bread (real world run throughs).

I'm doing my PhD looking at the fluid dynamics inside bioreactors (relevant eh!) and spend a lot of my time staring at large datasets, crying and failing to explain myself very well.

I was thinking about how all that frametime data could be squashed into a single number for each runthrough. Couldn't you take the mean of all of the successive differences in frametimes (frame durations)? For example, if five successive frames took 10ms, 11ms, 13ms, 9ms and 8ms to render, you could add up the magnitude of the duration differences - 10-11, 11-13, 13-9 and 9-8 - giving you 1+2+4+1 = 8ms. Dividing through by the number of intervals would give you 8ms/4intervals = 2ms/interval and that 2ms would be the mean duration difference between successive frames.

Cards which produced smooth 'natural' transitions in framerate as the scene changed in complexity would have very low mean frame duration difference, while cards with rapidly fluctuating frame durations would produce much larger numbers. You could express this number just as a time or as a proportion of the mean frametime - I'm not sure which is more important."

Tech discussion:

  • AMD's Radeon R9 290 graphics card reviewed - Read more

    Updated: Retail Radeon R9 290X cards may be slower than press samples - Read more

    Nvidia's GeForce GTX 780 Ti graphics card reviewed - Read more

    The SSD Endurance Experiment: 200TB update - Read more

That's all, folks! We'll see you on the next episode.TR

The TR Podcast bonus video: AMD, Zen, Fiji, and moreWith special guest David Kanter 53
The TR Podcast 175: The Zen of chipmaking and ARM's Cortex-A72 revealedPlus listener mail, Valve's mod mess-up, and more 3
The TR Podcast 174: PCIe SSDs and FreeSync displaysPlus corny jokes, new tablets, and lots of listener mail 12
The TR Podcast 173: Torquing the TitanPlus sexy mice, listener mail, and musings on VR hand tracking 4
The TR Podcast 172: An Apple on yer wrist and an Oculus on your faceA Cyril-less debut goes better than expected 6
The TR Podcast 171: Nvidia takes heat, Carrizo runs cool, and Fractal stays quietPlus listener mail and other stuff 3
The TR Podcast 170.5: The Heimlich maneuverSuperfish aggravates our trust issues, and G-Sync vs. FreeSync doesn't help much 9
The TR Podcast 170: What the kids put in their PCIe slots these daysPlus Dell's lovely Venue tablet, Spiderman comes back to Marvel, and much more 5

Tags: Podcast