DirkW

It really comes down to the looks, - and what that is worth to you.
 
I'm happy with mine, but when I got it, I was fully aware of what I'm doing (buying an overly expensive piece of hardware just for the looks - ok plus the waterblock on it as well). If we leave those two things out... there's no way I would have bought this board, as you can get about the same hardware for roughly half that price.
 
Anything ROG is pretty much always want, not need.

Should work, yes.
 
But are you sure the GPU is the lowest part of the loop (not only counting the obvious components like blocks, res, pump and rad but also of the whole tubing, etc.)? In my loop the lowest part is actually a section of tubing - and I put a T-connection in there, with one end being a dead end with a valve (and an additional plug just to be safe).

@IhazHedont
 
Unless I'm missing something on first glance, a loop like that wouldn't work well at all. The water would just rush straight through on the right side (least resistance) and not go through the blocks to cool the GPUs at all.
 
You'd have to move the top outlet on GPU #1 to the left port, so the water has to go through the block to get out, not just pass straight through the ports. And when you use 2 identical GPU blocks (otherwise don't even bother with parallel!), resistance should be be equal, and water should "divide up" and each half go through one particular block before exiting.

It really comes down to the looks, - and what that is worth to you.
 
I'm happy with mine, but when I got it, I was fully aware of what I'm doing (buying an overly expensive piece of hardware just for the looks - ok plus the waterblock on it as well). If we leave those two things out... there's no way I would have bought this board, as you can get about the same hardware for roughly half that price.
 
Anything ROG is pretty much always want, not need.

With rapid mode enabled, people usually get numbers in the 3000's and come asking why their SSD is so super fast...
 
@OP: It's really a bit mysterious, since the write speed doesn't seem to be affected at all, since it is still in the 500+ range. The 840 Evo had this read speed degradation for older files, which ld to a firmware update from Samsung, but I haven't heard anything like that about the 850 series...

Not from memory, but a quick google search led me to this:
 
http://www.overclock.net/t/1227636/how-to-change-sata-modes-after-windows-installation

You get exactly what you pay for. The disk has 1 TB=1000 GB. The manufacturer does everything 100% correct, nobody is being ripped off even a single byte.
 
Problem is Windows, while claiming to show GB, doesn't actually show GB (Gigabyte), but GiB (Gibibyte) instead - a slightly different unit, just wrongly labeled. So, instead of complaining about harddrive manufacturers, you should go and complain to Microsoft for making you see less space than you actually have. 

And, also sadly, they are way more expensive than normal Reds.

Just so that Tim Drake understands that it'll only work for certain combinations, but not in general.

Just so that Tim Drake understands that it'll only work for certain combinations, but not in general.

You get what you pay for - I doubt that you'll find something in that price range that can stream four 1080p streams simultaneously.
 
Personally I think Synology and QNAP have the best NAS systems out there, but like I said: for what you (apparently) want, you'll probably need to dig a little deeper into your pockets.
 
Something like a Synology DS415play or a QNAP TS-469L or so should work, but again: not for that price.

No. 1TB = 1000GB. And decimal is actually correct, since "tera", "giga", "mega" etc. are decimal unit prefixes.
 
The Problem is, that windows reports it in gibibyte instead of gigabyte (while still wrongly calling it gigabyte). So 1 TB = 931 GiB (not GB!).
 
As far as I know, Linux uses the decimal sizes and therefore shows drive sizes correctly, without causing this confusion.
 
So you've got to blame the OS/Software, not the hardware.

It is because most people on this forum are immature and ignorant. I usually come here for the excitement and drama lol. On OCN you dont get away with half the shit you can on here. Hell, if you swear you get warned. This site isnt moderated very well imo.

It's nether a driver issue nor anything that a firmware update could fix. It's the hardware config - and it's not a bug, it's been designed that way. The "only" problem is that nvidia didn't tell anybody until they got caught.
 
The card however, even if it does not quite have the specs it was supposed to have, will still perform exactly like all the reviews and benchmarks showed.

Just to get out of the way (in case it is not clear yet):
 
There are i7s of several generations and for several different sockets and chipsets.
 
So there is no i7 or Haswell, there are i7s (but also i5s, i3, etc.) from either 1st Gen (don't remember the codename(s) I'm afraid), Sandy Bridge, Ivy Bridge, Haswell (and soon Broadwell, Skylake, and Cannonlake, etc.) generations. Plus the versions for the enthusiast platforms (different sockets, more cores and PCIe lanes mostly).

Im gonna have to side with @Enderman on this one. Everyone took a look at the performance benchmarks and decided that for the amount of money they were spending this card made sense, and bought one.
 
Now a few months have past and a very few people have noticed some rather odd behavior of these cards and found that there are two partition of memory on the card and the second .5GB partition is much slower than the first 3.5GB. Sad face. The performance of your cards is no different from when you bought them and now all of the sudden people have convinced themselves based on this new information that their cards are no longer worth the money. I don't get people some days.
 
Now the second half of this story is where I do believe Nvidia does owe the consumers a bit. The 56 ROPS and 1792 KB L2 Cache vs the marketing claim of 64 ROPs and 2048 KB of L2 Cache. Although I do not believe that Nvidia was deliberately trying to deceive customers, as I have first hand experience with the occasional inept capabilities of marketing departments, however this mistake did somehow make it out to the public. This is where I do think Nvidia should offer a small refund ($15-$20 maybe) for their mistake or a choice of a couple of games on Steam. This error is in my opinion the only true error that Nvidia made with this product, the marketing information.

*sigh*
 
Maybe you should just spell it out clearly instead of posting a picture again and again...
 
What people call the lie here is not the 3.5GB vs. 4GB of VRAM. It's the 56 vs. 64 ROPs and 1792KB vs. 2048 KB of L2-Cache. These are things you do not get (unlike the 4GB of VRAM).
 
Does it suck what nVidia did there? Yes. A lot.
Is the GTX970 still a good card that performs just like in the reviews and in all the months people were enjoying it now? Yes, sure still does.
Is the latter what the outcry is about? Nope. Not at all.
 
So yes, the card's performance did not change at all. But the specs, nVidia said it would have, did. So, 2 very different things to judge here:
 
1. People who got this card because of its performance in benchmarks and reviews, actually got what they paid for. The card still delivers just that.
2. People who got the card because of its specs... well, they were screwed over by nVidia. (Not that I believe that the 2nd group is more than a minority, though).
 
Will it hurt nVidia in the long term? Don't think so. For many people it's already the evil empire (just like Intel, Microsoft, Google, etc.), so one more bad thing... who cares? People actually kinda expect this shit from some companies.
 
Unfortunately that's how the world goes... it will be forgotten soon by 99% of people.
 
I will certainly not forget it, but my personal experiences with products from both companies will affect me more (and GX970 is no personal experience for me), so it's actually kinda likely that I will ignore it and buy nVidia again in the future...

OS X ≠  IOS

The bottom slot on my board is only PCIe 2.0 4x, and isn't even connected directly to the CPU but to the chipset - IMHO not a place to put a GPU.

That anything stated by a tech youtuber is 100% factual no matter what, and cannot be disputed. to do so is blasphemy. 

In fact way too fast to be true.

Neither do mine. And yes, cache would be the only explanation (as can be seen regularly in the astonished reports from people using Samsung's Rapid Mode on their SSDs ) - but then again it's technically no longer th speed of the drive itself.

Maybe it's about giving you options, not making you use them all at once? :rolleyes:

Read is actually not limited, and also the write limitation is much more than most people think. With a modern, quality SSD and normal usage you will not get into these limits for years.

Read is actually not limited, and also the write limitation is much more than most people think. With a modern, quality SSD and normal usage you will not get into these limits for years.