Slashdot Mirror

The write cache on the Crucial M500 survived my power plug testing with similar properties to this article, but it doesn't have SMART data for drive wear. Can't take that one seriously for business use. I've been very happy with the figures I've seen from Intel drives on their internal lifespan tracking; see my Intel SSD lifespan for example. I'm not the only one who noticed this flaw in the M500, the Tech Report review has another complaint.

None of the cheap and easy to buy Samsung drives (840 and 840 Pro) claim any power loss protection, and they all fail this sort of test. They have enterprise models that might work, but those wouldn't fit within the budget parameters here. I find it hard to take those seriously when the 840 drive they share design features with are so terrible handling lifespan failures.

I'm not aware of any Sandisk models with that feature, but I haven't gone looking for them either.

The only drive really missing here that might have passed are the Seagate 600 Pro models. Those haven't been shipping long enough for me to recommend them yet.

Samsung doesn't claim any power loss protection even in the 840 Pro models, and as you've seen it shows in the real world. The 840 models are fast and cheap, but if you care at all about longevity they're just unacceptable. It was no surprise to find the 840 was the first to fail in the largest longevity test I know of.

The power loss protection on the Crucial M500 worked fine in my initial tests. It can't be taken seriously as a reliable drive because it doesn't have any SMART data on longevity. There's no way to know when the drive is wearing out, so it's pretty much useless for serious work. The one I bought for testing is in my laptop, it's a fine drive there. See Tech Report for a review complaining about the missing SMART data, I'm not the only one who noticed.

Intel's data on wear is very good, see my look at the 320 vs. 710 lifetime for example. The replacement models, DCS3500 and DCS3700, are even better drives in every way.

...so unless you think the government has a secret back door into every encryption algorithm...

Where have you been man? See here.

There's a big difference between a backdoor in a published encryption algorithm and a backdoor in commercial encryption software/hardware. It's much harder to hide a backdoor in a well known algorithm that's been under international scrutiny. Though I do have my doubts about the ECC constants

Do you think the government should be able to retrieve your private conversations on an analyst's "hunch"?

Not at all, I think the whole thing is total bullshit, beyond expression and pushes the bounds of humanity itself. However That means nothing to anyone but me, maybe some others. What our government needs is some small, no matter how small, reason to point a finger. Don't forget, the whole war in Iraq started because of a "hunch". And they found out that that hunch was wrong. Well... the wars (yes wars) machine keeps on turning, brother.

...so unless you think the government has a secret back door into every encryption algorithm...

Where have you been man? See here.

For example, the Intel SSDs we use are rated to withstand 100% of the drive's capacity in writes every 24 hours for many years on end. A consumer drive couldn't do that for more than a few weeks, perhaps a month or two.

Every drive tested here has handled over 1000x capacity of writes, which is around 3 years at your benchmark of 100% capacity per day. Every one of those drives is a consumer drive. Some are showing signs of eventually failing, but none has lost one byte of data.

In real-world usage, that means consumer SSDs are easily good for 10 years before write endurance becomes a problem.

Oh, no doubt. It's just that Nehalem was a bit too soon (and I was more broke then) and afterwards it just didn't seem that exciting - I thought I'd do it with Haswell, what with the TSX, AVX2 and what not, but then the -K series got crippled for differentiation reasons, and the heat issues, and overall focus on energy over performance and just... meh.

And while I'm sure the processor is limiting the graphics card somewhat, as far as I could tell, it's not that much.
Let's say my overclocked Q6600 is roughly as fast as the AMD A10-5800K (good job, AMD!):
http://anandtech.com/bench/product/675?vs=53
The A10, while being the slowest of the bunch, delivers just 10 fewer fps in Crysis than the 4770K:
http://techreport.com/review/24879/intel-core-i7-4770k-and-4950hq-haswell-processors-reviewed/9
yeah it's worse in the worst case but again, just meh.

But I am going to do it next year for sure, so I hope Nvidia is working hard on Maxwell as well!

As the title says, I recently stumbled upon this site.

http://techreport.com/

The reviews seem well thought out and complete with the good and the bad. They are not user reviews, but I am finding them very helpful as a starting point.

I have n connection with them, but so far I am impressed. The long term report on SSDs was particularly good since I have several and was wondering about life claims by the various manufacturers.

They don't cover everything, but for now it is the first place I look to see if there is a review.

You know, like this relentless drive to make things flat, single shaded and without borders / lines or dividers making information / data easier to identify. It's all mashed together.

Furthermore the tabletification of web pages, the urge to push Google+ on their customers, really importantly also is the move to remove the menu (context) button on Android devices instead for a multi-task button. One of the _primary_ reasons, literally one of the biggest factors in me moving to Android from Apple was this genius concept of the options / settings / context "other" menu button ALWAYS being in a consistent spot and not fucking floating around on the screen randomly like it does on an iphone! (Yes, I know Blackberry did it first and credit to them too)
But nope, let's change that idea for a multi-task button, what the hell? No thanks - I'd rather just do a double tap of home for multi-task.

Then there's the removal of the text on icons - yes, I get this allows them to be able to regionalise things MUCH MUCH faster, can't deny that benefit to THEM but to me? To identify these ambiguous looking icons? Sure some are consistent, like a green telephone icon at a diagonal angle for call and flat 'hanging up' red phone for hanging up but some of the other stuff is not so clear (and even if you figure it out it takes time to process) It took me at least 30 seconds to figure out how to edit a draft email in the new gmail app about 6 months ago, because they removed the fucking text 'edit' - there's a small pencil icon instead, tucked in a corner.

So yes, a poor UI design program is bloody essential, if not the fucking firing of developers when a project is complete. STOP fiddling, STOP justifying your jobs somehow. Some internet websites / apps are services now. They are like water, it's like a toaster, does it REALLY need to actually change?
Do I really need multiple inboxes? I LIKE the priority inbox, now already the damn thing is 'grandfathered' for their new design, how long until that multi-inbox thing is forced on me as the old one is finally terminated / not supported?

Sorry to rant but Google for the past 18 months is NOT the google of previous times, they are driving me frankly, fucking bonkers. Get your shit together.

http://i.imgur.com/h7Mpdq4.gif
See also:
http://techreport.com/blog/25657/google-i-love-you-but-youre-bringing-me-down

The FPS per dollar scatter plot on page 9 of the linked article (here) is really telling. There's a surprisingly tight correlation between dollars and FPS for almost all of the cards, and then the GTX Titan is way off in no man's land. Nvidia's going to have to drop the price, unless it's just there to soak up money from people with more dollars than sense.

You're quoting reviews that are months old. The newer driver updates were designed specifically to fix these problems, and for the most part, they have succeeded. (There are still issues in some specific CrossFire and/or multi-monitor configurations, but these won't affect most users.)

One of the reviews you cited was from The Tech Report, which did a good job of documenting these frame pacing issues with hard numbers a couple of months back. Well, let's see what they have to say about the R9 290X now:

You can see from the raw plots that the 290X looks good, with more frames produced and generally lower frame rendering times than anything else we tested. Every card encounters a few slowdowns, and the spikes on the 290X aren't anything exceptional.

Our "badness" index concentrates on those frames that take a long time to produce. For the first two thresholds of 50 and 33 ms, the results are pretty similar among the newer GPUs, which again suggests a CPU bottleneck or the like. However, for slinging out frames 60 times per second, once every 16.7 milliseconds, the R9 290X is easily the best choice.

Cant seem to find the original nvidia vs ati render stuttering, but these will do.
http://www.anandtech.com/show/6857/amd-stuttering-issues-driver-roadmap-fraps
http://techreport.com/review/24022/does-the-radeon-hd-7950-stumble-in-windows-8/10

I couldn't care less if this this the cheapest/fastest card on the planet.
Until AMD fix the core stuttering issues with their drivers, instead of just patching it for a AAA game now and then. I'am really not interested.

Frame rate isnt everything, stability and consistent render times of those frames are. Yes, i have an ATI card and i'am regretting it.

A solution you can pick up for 99 bux at Walmart since 2009 apparently - http://techreport.com/news/16466/marvell-unveils-99-wall-wart-linux-pc

There's a new SSD twist to this class of problem popping up more lately too. When SSDs wear out, they will sometimes quietly lose data. People doing SSD endurance testing with Anvil's tool stop the system periodically to see if it still works after being powered off for a bit. That's how some overused SSDs die at the end, and it's seriously ugly.

These chips are slightly faster (given equal core counts) than their predecessors but not in any interesting way.

  However, you have to remember that these are really server chips that are repurposed for high-end desktop use. The one vital metric where these chips shine is in their power consumption (or lack thereof): Techreport did a test where the 6-core 4960X running full-bore is using about the same amount of power as a desktop A10-6800K part ( http://techreport.com/review/25293/intel-core-i7-4960x-processor-reviewed/9 )

That level of power efficiency will do wonders in the server world and these chips (and their 12-core bigger brothers) should do quite well in servers.

Comparing most expensive chips isn't fair or useful. Intel's most expensive chips can cost a lot more because AMD doesn't have anything competitive.

A AMD FX-8350 costs $200. In Intel land, a i5-3750 is the right cost equal, at about $215. Intel's lead is so large that even a previous generation unit from their line up is approximately equal performance to AMD's current models. Which of those two is faster depends on the benchmark.

At the $100 end of the market, there are a few really cheap models where AMD has a price performance lead over Intel. But the minute you get to even $200, they are at best evenly matched. And Intel's built in HD graphics chips are getting better fast enough that even the AMD APU models won't have a lead at any price level for much longer.

FWIW I'm "in the know" and this was completely unsolicited. http://techreport.com/forums/viewtopic.php?f=1&t=81535

ermm, did you get this data from ms financial reports or outta your ass? because its bullshit. only exchange, windows, and office make money at redmond. something that makes SOME money after you invest TONS of money (and continually requires more money than it makes back) is still losing money hand over fist. and dont get me started on the hardware division (zune, kin, bob, keyboards and mice, etc).

Hehe, you didn't even bother to check that it was actually true before spouting ill-informed nonsense?

Maybe they just need to rename the benchmarking binary (I vote quack.exe)?

I jest, but only partly. You see, there is truly nothing new under the sun.

This.

Does anyone honestly think anyone is not optimising their product for the products favourite benchmarking programs? Does anyone honestly think they aren't using every dirty trick in the book to get a few points more?

I have no trouble believing that Samsung has optimised their phones for benchmarking, the same as I have no problem believing that Apple optimises their phones for benchmarking (Apple would be even worse, blocking benchmark applications that dont give them favourable results). The only difference is that someone is calling out Samsung.

ATI, Nvidia, Intel, IBM, Apple, HTC... Everyone is doing it because you (the general you, no the parent) have made these benchmarks important to decision making.

Also it's trivial to over clock an android phone. So an easy test for this would be to get an SGS4 and sert it to 480 MHz, test it then OC it to 532 MHz, test it again and compare both results to a non OC'd SGS4.

Maybe they just need to rename the benchmarking binary (I vote quack.exe)?

I jest, but only partly. You see, there is truly nothing new under the sun.

I've been all AMD almost forever, for this reason among others.

http://forums.pcper.com/showthread.php?470102-Intel-s-compiler-cripples-code-on-AMD-and-VIA-chips 2010

http://www.theregister.co.uk/2009/12/16/intel_ftc/ 2009

http://techreport.com/news/8547/does-intel-compiler-cripple-amd-performance 2005

I found those three on the first page of my search results, and quit looking. Different search terms and a more determined search will find hits as old as about 1999, maybe even older. Hard to remember, but I think I first became aware of compiler cheats by Intel around 2000 or 2001. Prior to that, I naively thought that a compiler was a compiler.

Not only that, but the new Mac Pro is probably the most original desktop computer design since.. desktop computers were invented.

Look, I *own* a Mac Pro. It's a great machine. The new one... it's neither innovative or great.

A machine that requires external expansion? Not new since Sinclair ZX-80 (or probably earlier.) A high power design that uses a vertical cooling tower? Audio's been doing this forever; all kinds of actually innovative cooling designs can be found in the history of audio power amplifiers. Tunnels, polygons, forced air, liquid, even semiconductor. Odd physical configuration? Tons of 'em out there. Raspberry pi. Mini ITX. or this. The new Mac Pro has a new CPU in it, as does virtually every iteration of these machines; it has a flash boot drive, like a lot of computers; and you're gonna have to spend a lot more money to make a worthwhile computer out of it... like a ZX-80. It's also likely to be very, very expensive. You know what that spells? "Not Buying"

But hey. You can always by a Mac Mini or an iMac. Those, at least, are working multicore computers out of the box. Or, I hasten to add, the current generation Mac Pro, which is a great machine.

FIXED LINK

We do love our big numbers, but there are limits to what our eyes can perceive in FPS. What does this mean for real world applications like video encoding and password cracking? How long do we anticipate having to wait for tech like this to get affordable? Also, how does this compare to the nVidia Tesla, the current gold standard in password cracking?

I saw only one reference to nVidia Tesla (and no references to password cracking or video encoding) in those reviews (@Tech Report), and it might be damning:

Speaking of things that don't matter much, Nvidia has decided to scale back the GTX 780's capacity for double-precision floating-point math. Double-precision support is built into the GK110 GPU because of the chip's compute-focused role aboard Nvidia's Tesla products. Real-time graphics basically don't require that level of precision.

NVIDIA has already announced stacked RAM for their future Volta cards (first link I could find).

..and by 'trounced' you mean 'not that big of a difference but I'm going to pretend', right?

It looks like the Blue takes about 20% longer on random I/O, based on these tests. Writes are even worse, at 30-40% longer.

Then, jump to page 8 on those tests, and see that the Black gets 40-60% more IOPS on every workload.

The Frame latencies by percentile graph they create now is the right way to look at this data. It's a sort of probability distribution function for slow frames. Nothing simpler will capture the complexity of the problem. You can't usefully boil the universe of rendering latency issues into any single number.

The worst frame will vary based on card and game, and the tools available to reviewers are not practical to find them. And what this debacle has shown is that even though they're limited, the tools being used by reviewers are sometimes better than internal QA at the manufacturers.

The root kit? Which one? The CD rootkit or the USB drive rootkit? http://techreport.com/news/13096/sony-usb-thumb-drives-come-with-rootkit

Also, having just looked at the videos, I actually find the AMD version to be more visually pleasing - yes, there's a big "jump" at a few spots in the video, but it's in no way regular, while the nVidia jumps (albeit a smaller distance) periodically - roughly once a second. It's like a rhythmic pulsing... really weird.

And I can't say I've seen this issue on my HD7750 (awesome little card - no PEG connector needed, doesn't break 50C running Starcraft 2 on high, low power consumption)... although that might be because I'm not running the beta driver.

Isn't it also an AMD CPU issue? Or is the problem simply much more pronounced with this particular line of Radeon GPUs?

Or was the CPU problem long corrected?

This article from TechReport in August convinced me not to go AMD for my next gaming PC:

http://techreport.com/review/23246/inside-the-second-gaming-performance-with-today-cpus

Seriously, WTF. He was drawing "weapons" in his notebook.
Was he plotting out detailed plans for causing mass destruction? No?

Maybe he drew a detailed blueprint for an atomic bomb? Or maybe just an IDE? No?
TFA doesn't say. Another article quotes his mom saying that he drew a glove shooting fire. A glove... shooting magical fire... WTF?
Ok maybe it was this one. Maybe he was planning to build his own? That would have been aweso... err I mean criminal!

But wait, since the staff called the police then there must be some threat that he poses, right? Let's search his house just to be safe. Oh look! We found scary-looking exposed electronic parts!
Oh, and some "chemicals" that could be mixed together to make an explosive!
Let's throw the book at him! Great job here, boys. Let's grab a round and celebrate.

TechReport analysed the nVidia 680 a bit after its release and had a piece on adaptive vsync which should answer your question.

Quoted from an nVidia software engineer:

There are two definitions for triple buffering. One applies to OGL and the other to DX. Adaptive v-sync provides benefits in terms of power savings and smoothness relative to both.

- Triple buffering solutions require more frame-buffer memory than double buffering, which can be a problem at high resolutions.

- Triple buffering is an application choice (no driver override in DX) and is not frequently supported.

- OGL triple buffering: The GPU renders frames as fast as it can (equivalent to v-sync off) and the most recently completed frame is display at the next v-sync. This means you get tear-free rendering, but entire frames are affectively dropped (never displayed) so smoothness is severely compromised and the effective time interval between successive displayed frames can vary by a factor of two. Measuring fps in this case will return the v-sync off frame rate which is meaningless when some frames are not displayed (can you be sure they were actually rendered?). To summarize- this implementation combines high power consumption and uneven motion sampling for a poor user experience.

- DX triple buffering is the same as double buffering but with three back buffers which allows the GPU to render two frames before stalling for display to complete scanout of the oldest frame. The resulting behavior is the same as adaptive vsync (or regular double-buffered v-sync=on) for frame rates above 60Hz, so power and smoothness are ok. It's a different story when the frame rate drops below 60 though. Below 60Hz this solution will run faster than 30Hz (i.e. better than regular double buffered v-sync=on) because successive frames will display after either 1 or 2 v-blank intervals. This results in better average frame rates, but the samples are uneven and smoothness is compromised.

- Adaptive vsync is smooth below 60Hz (even samples) and uses less power above 60Hz.

- Triple buffering adds 50% more latency to the rendering pipeline. This is particularly problematic below 60fps. Adaptive vsync adds no latency.

So triple buffering is bad because it could cause an intermediary frame to be dropped, resulting in a small visual stutter despite being 60fps. There's a video of adaptive vsync on YouTube.

Anyone that actually knows anything about the GPU industry knows that both AMD and NVIDIA graphics suffer from these latency spikes, but it's not with all their SKUs. NVIDIA's 660 Ti works well in this case, but their 670 and 680 has more latency spikes than the competitive AMD cards do. The 7850 demonstrated here is an anomaly for AMD. None of their other cards do this. Look at past reviews from Techreport and you will see what I mean.

Sorry, but ATI suffer from this far more than Nvidia as apparently it is something nvidia actively try to improve:

http://techreport.com/review/21516/inside-the-second-a-new-look-at-game-benchmarking

You have to read the full article, but even though it is old now it sound like it is still relevant.

Well, you have that information from the graphs of the X% of frames under latency Y, but it's a lot of information to report not in a graphic form :) Most site I`ve seen using this `new` measurement display 'extracts' of frame rendering time vs time in harder locations, a graph plotting the frame rendering time (Y axis) vs the frame ID (X axis), which will provide you with the framerate, total number of frame renderer, average latency, maximum latency, etc, and the finally provide the analysis of this information with the graph I was talking of initially.

Here's a sample: http://techreport.com/review/24022/does-the-radeon-hd-7950-stumble-in-windows-8/8

However, if you actually read the Tech Report's review of the GTX 670, you will find they say the exact opposite. The GTX 670 has ridiculously low latency compared to the Radeon 6990 and just a bit lower than the 7950 and 7990.

As clearly seen on page 3 of the 670's review.

Look at the bottom Skyrim graph in
http://techreport.com/review/24022/does-the-radeon-hd-7950-stumble-in-windows-8/8

The slow frames always follow extra-fast ones. Looks like some work is being deferred past a frame boundary?!

> but IP itself in and of itself is not an obstruction to progress.

You want to try to tell that to Carmack who has an algorithm named after him (Carmack's Reverse) that he independently invented and he CAN'T use it due to idiotic IP laws.

* Description of his notes independently (re)discovering the algorithm
http://wayback.archive.org/web/jsp/Interstitial.jsp?seconds=5&date=1233022175000&url=http%3A%2F%2Fdeveloper.nvidia.com%2Fattach%2F6832&target=http%3A%2F%2Fweb.archive.org%2Fweb%2F20090127020935%2Fhttp%3A%2F%2Fdeveloper.nvidia.com%2Fattach%2F6832

The patent in question:
* http://www.google.com/patents/US6384822

Quoting John Carmack:
* http://techreport.com/news/7113/creative-patents-carmack-reverse

"The patent situation well and truly sucks.

We were prepared to use a two-pass algorithm that gave equivalent results at a speed hit, but we negotiated the deal with Creative so that we were able to use the zfail method without having to actually pay any cash. It was tempting to take a stand and say that our products were never going to use any advanced Creative/3dlabs products because of their position on patenting gaming software algorithms, but that would only have hurt the users. "

So again, what are you smoking by claiming "IP itself in and of itself is not an obstruction to progress."

"Can you actually explain what the advantages are of having that extra 10% ? Once you go above 60 FPS you don't get the "stutter" and are actually rendering duplicate frames. So what is the Net gain of additional frames especially if your monitor (60 hz, or 120 hz) can't display them ?"

First, 60 FPS is an average...

Second, read the following article, http://techreport.com/review/22890/nvidia-geforce-gtx-690-graphics-card/3

Make a RAM Drive! Try one of these products and insert your RAM. Then you have the FASTEST Hard Drive in the world! http://www.tomshardware.com/reviews/hyperos-dram-hard-drive-block,1186.html http://techreport.com/review/9312/gigabyte-i-ram-storage-device

Yeah, thanks for making me feel bad for breaking not one but two of my nice 1920x1200 LCDs.

OTOH, 1920x1080 is getting cheap enough that you could grab 2 or even 3 for the price of one WUXGA display. Which makes me want to work and/or play three screens...
  http://techreport.com/review/23217/triple-screen-gaming-on-today-graphics-cards

But since I'm a cheapskate, I just picked up a handful of cheap 19" - 21" CRTs from craigslist for between $5 - $20 each.

For laptops, I would just as soon try to set up compiz-fusion to scale (with full anti-aliasing) a large VNC session or something, so I can zoom in and out of a large X server session. I'm kinda wondering why more UIs aren't really going this route (other than maybe being slightly nauseating.)

IBM PCs of the era had a similar option: attach the RAM to the ISA bus via an add-on card. Like the Amiga (and most computers of that era), the expansion bus was the processor bus (with a bit of buffering and maybe a tad bit of glue logic, but not much more).

As processor speeds increased, this became a problem. Many peripherals just weren't designed for the increased speed, so they divorced the bus speed from the processor speed by making it a fraction of the processor speed (ISA) or going asynchronous (Amiga Zorro III). This became quite pronounced with PCI (max 66MHz, even if you're running a 3.0GHz CPU); you can add memory onto the bus, but it will slow you down if you try to use it as main memory.

That doesn't mean it can't be used at all these days. The cluster computer folks have a concept called NUMA, or non-uniform memory access, where memory isn't considered necessarily equal in speed. Or you could treat it like a very fast SATA drive, provided you have the necessary means of keeping power to it during power failure events (or use it only as temp or swap space).

I also heard about Nvidia engineers "helping out" with code. Result is usually some spaghetti garbage that only works good on Nvidia.
For example Crysis 2 tessellation "optimized" to run smooth on Nvidia cards that excel at pointless tessellation.
http://techreport.com/review/21404/crysis-2-tessellation-too-much-of-a-good-thing/2

I'm confused. When Vista came out with their Aero effects, it was a massive CPU and memory hog in addition to requiring GPU power. Compiz on the other hand ran on very basic hardware and did far more.

In Linux I have more control over font rendering and sub-pixel hinting.

In what way has Linux desktop graphics lagged at all? Mac OS X and Linux have led the way while Windows has followed poorly. And in case you haven't been paying attention, Windows 8 is coming out. It is 2012, and they still haven't figured out how to scale down to small displays, and scale up to very high resolution displays.

http://techreport.com/review/23631/how-windows-8-scaling-fails-on-high-ppi-displays

That article flies in the face of AMD's recent statements regarding Steamroller where they claim they are moving from doing layout by hand to maximize performance and density as practiced in Bulldozer to using a high-density cell library for layout.

Shown above is a portion of the chip's FPU. The top image comes from a current Bulldozer chip, which employs the hand-drawn custom logic that's generally used in high-end x86 CPUs. The lower image comes from a potential future chip that uses a more automated high-density cell library. On the same 32-nm process node, the high-density library purportedly crams the same logic into 30% less area, with 30% less power use. As the slide notes, gains on this order would usually come from the transition to a newer, smaller fabrication process. We'd expect the more automated approach to design to reduce AMD's time to market, as well.

AMD allowed websites to publish a preview of the benchmarks before the estimated date if they only focused on graphics performance. This is an unfair move by AMD.

Read http://techreport.com/blog/23638/amd-attempts-to-shape-review-content-with-staged-release-of-info for more details

(maybe in a couple of weeks you will find that AMD Trinity APUs have abysmal x86 performance compared to Intel CPUs)

Disclaimer: I own a laptop with an AMD cpu inside

Ohhhh prior art

http://techreport.com/gallery/index.x?id=23432&image=59639

I'm sure I've seen those light switches before.

Valve's Gabe Newell calls Win8 a 'catastrophe,' wants Linux to thrive

The Valve founder started his response by saying Valve owes its success to the inherent openness of the PC as a platform, but going forward, the company will need to take an active part in "[making] sure there are open platforms."

http://techreport.com/discussions.x/23315

The primary reason a 48fps average may not be good enough for gaming is because of variability. In a 48fps movie, every single frame comes 20.83ms after the one before it. If a game has an average of 61 FPS, it could be that sixty frame times are 15ms and the sixty-first is a very noticeable and jarring 100ms. Hardware reviewers, starting with Tech Report, are starting to catch on to the fact that the longest frame times matter much more than the average.

Two other reasons why a framerate where movies are smooth may not suffice for gaming:

• Normal 3d rendered frames are like using a camera with an infinitely fast shutter. The difference between the resulting sudden motions and the natural blur in frames that have had exposure times comparable to their frame display times is the same thing as the difference between "jaggies" and smoothly antialiased edges in a single frame, except in the time domain rather than the spatial. Artificial motion blur effects try to improve on that, but often it's easier to just increase framerate (just as increasing the resolution makes the "jaggies" in a single frame less obvious).
• The relationship between Input lag and frame delay in a game is somewhat complex. It would of course be possible to have a thousand or more frames per second and still have a full second of input lag, and it's possible to have input lag low enough for twitch FPSes with less than 60fps. But if a game's engine design, buffering, and the hardware input and output pathways were such that input lag is 3*frame time + 50ms, you would notice improvements in input lag from having ridiculously high frame rates well beyond the point of smooth motion.

Here's the study showing little difference between 30fps and 60fps for a first person shooter. Note the dramatic differences up to 15fps, where we begin to perceive many things as motion rather than individual frames, the still-quite-noticeable improvement in moving to a 30fps rate, where quick full-frame motions are starting to appear smooth, and the much smaller change (esp. the almost-zero perceived quality change) between 30fps and 60fps. Since the frame time difference between 30fps and 48fps is 12.5ms and the difference between 48fps and 60fps is only 4.17ms, I sincerely doubt their test setup would have showed any difference between 48fps and any higher rate (even 240+). However, their test setup was different from most games in one or maybe two of the three points I mentioned above. The frame rate limitations in the paper were artificial- the computer was capable of producing a much much higher framerate- so the frame times were probably almost exactly consistent. It's also possible that their mechanism for limiting framerate didn't give as strong a link between framerate and input lag as most games have (you'd have to look at the code to be sure).

http://techreport.com/discussions.x/22366

The Tech Report has chimed in with its own review, which contains a unique look at gaming performance with the integrated graphics and discrete GPUs. There's also a dedicated overclocking article that looks at the experience on four different motherboards.