All these fire sales on high-end Nvidia video cards...

[Chalenge]

No, Ripley, I told you what I do everyday. Sorry you choose not to accept it. You should probably stop giving advice now.

[MrRiplEy[H]]

No, Ripley, I told you what I do everyday. Sorry you choose not to accept it. You should probably stop giving advice now.

You should probably stop being so arrogant and think everyone uses computer like you do. 100% sure even you do not render 24 hours per day as you tried falsely to present. This discussion is on a general level, it's not about you.

[Chalenge]

Quite to the contrary Ripley. You have basically just confirmed that it is possible to kill an SSD rather quickly. The problem is that most people never ask how someone might use the drive and like you just tell someone to go ahead and get an SSD for some ripping speed! I like how you threw the arrogance thing in there, though. That's cute. People do that here when they got nothing, and you got nothing.

Now, unlike you I just showed everyone how to find out what their data usage is and how long one of these drives will last in their particular situation. I can tell you, for instance, that the longest lasting SSD for my situation would go for 27 years, but it isn't big enough to hold the programs I use. In my situation I need a drive that is at least 640 GB. With that in mind and checking the data again I see that an SSD will last about one year, at best. Because that hard drive will cost more than $400 I decide I will buy an Enterprise level HDD. The slated lifespan of a WD Re drive is 6 years in the median term, while the longest lasting 640 GB SSD is six years if all I do is cruise the Internet and answer a few forums posts. WD Re 4TB = $275. 1 TB SSD = $459.

I suppose my arrogance in deciding to share that information is so offensive I should stop, huh? Instead, why don't you stick to the facts.

[MrRiplEy[H]]

Quite to the contrary Ripley. You have basically just confirmed that it is possible to kill an SSD rather quickly. The problem is that most people never ask how someone might use the drive and like you just tell someone to go ahead and get an SSD for some ripping speed! I like how you threw the arrogance thing in there, though. That's cute. People do that here when they got nothing, and you got nothing.

Now, unlike you I just showed everyone how to find out what their data usage is and how long one of these drives will last in their particular situation. I can tell you, for instance, that the longest lasting SSD for my situation would go for 27 years, but it isn't big enough to hold the programs I use. In my situation I need a drive that is at least 640 GB. With that in mind and checking the data again I see that an SSD will last about one year, at best. Because that hard drive will cost more than $400 I decide I will buy an Enterprise level HDD. The slated lifespan of a WD Re drive is 6 years in the median term, while the longest lasting 640 GB SSD is six years if all I do is cruise the Internet and answer a few forums posts. WD Re 4TB = $275. 1 TB SSD = $459.

I suppose my arrogance in deciding to share that information is so offensive I should stop, huh? Instead, why don't you stick to the facts.

LOL you pull numbers out of your behind again. The 640Gb SSD will outlive you and everyone on this forum if all you do is surf and play games with it. You got caught cheating, admit it like a man and move on.

You tried falsifying statistics on multiple occasions: First you tried posting the site with a 200Gb daily writes as baseline. The average computer user has average writes of 8-10 gigs/day, not 200 gigs. Then you manufactured a 'worst case scenario' usage statistic from your computer by pushing i/o to the max for 8 minutes and then pretending like your computer crunches 5 terabytes per day. Second lie. Then your statistic showed an estimate of mere 20Tb for the life of an SSD. Third lie.

For a general user write endurancy will never become a problem. I highly doubt it would become an issue to you either. Go ahead and post a 1 week statistic of your write/reads. If they continue showing 5Tb a day then I will believe you. Your usage patterns are atypical anyway, not something you should ever bring into discussion. Did I bring into the discussion that I get maybe 1 gig a day writes on average in my own use (on my gaming machine)? Or I don't even boot my game machine every day? No. There's no point discussing anything but the typical usage scenario.

[Chalenge]

No, Ripley. There are actual utilities available that I used to make this determination. Every one of them says the same thing. If I switched out the SSDs for the HDDs I have now the best I could hope for is six years, but the SSD of the proper size that I need would not live long. I am testing based upon my usage, not the 'standard user.'  Depending on 'standard' figures would be stupid.

Here is a run that I did today for 16 hours:



Now if I did as you suggested (you suggested I fudged the numbers), then the daily usage would have dropped more significantly than it has. However, what I did instead is stop the test before I commit to another render. After a second render the writes/day would go up, and the lifespan would drop. Now, according to your theory my hard drive would have died more than a year ago. Yet,. . . not so much.

Your advice is costing people money whether you choose to accept it, or not.

[MrRiplEy[H]]

No, Ripley. There are actual utilities available that I used to make this determination. Every one of them says the same thing. If I switched out the SSDs for the HDDs I have now the best I could hope for is six years, but the SSD of the proper size that I need would not live long. I am testing based upon my usage, not the 'standard user.'  Depending on 'standard' figures would be stupid.

Here is a run that I did today for 16 hours:

(Image removed from quote.)

Now if I did as you suggested (you suggested I fudged the numbers), then the daily usage would have dropped more significantly than it has. However, what I did instead is stop the test before I commit to another render. After a second render the writes/day would go up, and the lifespan would drop. Now, according to your theory my hard drive would have died more than a year ago. Yet,. . . not so much.

Your advice is costing people money whether you choose to accept it, or not.

Ahh yeah so you went from 5 terabytes to 141 gigabytes and still 8 hours of idling per day to go (or was part of that 16 hours night time?) so that would make it 94 gigs of writes per day in reality. If we mirror that to the tested write endurance of cheap SSDs (700 terabytes) you end up with a 7000 day life. That's 20 years. The more expensive drives you should use for work last past 1 petabyte writes.

You think 5tb vs 94 gigabytes is not that much different? It was only 53 (fifty-three) times higher. And this is without counting weekends and holidays when you obviously don't work lol (which would show in the 1 week or 1 month statistic I asked for true usage pattern). You can see your workload is in reality read heavy, your write/read ratio is only 0,079 and that's an extremely low ratio. Your usage pattern is actually just perfect for a SSD.

[Chalenge]

You're fantasizing too much about this Ripley. The lifespan of the drive at the point I tested it last had gone from 4 days to just under half a month. The first test was done when moving the 5TB project files into place. As the project changes it happens again, and the lifespan drops again to less than 3 days.

And now you're the one pulling numbers out of thin air.

[MrRiplEy[H]]

You're fantasizing too much about this Ripley. The lifespan of the drive at the point I tested it last had gone from 4 days to just under half a month. The first test was done when moving the 5TB project files into place. As the project changes it happens again, and the lifespan drops again to less than 3 days.

And now you're the one pulling numbers out of thin air.

Which drive are you using as your reference point? To my knowledge outside of USB flash drives there are no solid state drives that have a mere 20Tb write endurancy. Is it some 16Gb model perhaps? LOL!

I just presented you with a link to a test that shows that even the cheapest SSDs last up to 700Tb writes. Where did you pull your imaginary 20Tb model?

I'm not pulling numbers out of thin air like you are: http://techreport.com/review/27062/the-ssd-endurance-experiment-only-two-remain-after-1-5pb

YOU WON'T BELIEVE HOW MUCH DATA can be written to modern SSDs. No, seriously. Our ongoing SSD Endurance Experiment has demonstrated that some consumer-grade drives can withstand over a petabyte of writes before burning out. That's a hyperbole-worthy total for a class of products typically rated to survive only a few hundred terabytes at most.

[Chalenge]

Ripley, the entire point is that SSDs should be selected based upon specific usage trends for any particular user. As I have told you, and that you refuse to accept, is that there are users that you will run into that have usage patterns outside of your experience. There are usage models that exist outside of the capabilities of the larger SSDs to handle. Therefore it is cost prohibitive to have both an extreme usage model and a large capacity SSD. That's not the end of problematic environments for SSDs, either. SSDs are simply not ready, yet, for every potential usage model that exists, not for every potential environment. Therefore, you need to start asking more questions before making any further SSD recommendations.

You won't, but there it is. Now you can go back to insults and repeating exhaustive argument patterns, but I am going to leave it there. Instead of replying for another three pages all you have to do is say 'okay' and you will get the last word (as is your habit).

[MrRiplEy[H]]

Ripley, the entire point is that SSDs should be selected based upon specific usage trends for any particular user. As I have told you, and that you refuse to accept, is that there are users that you will run into that have usage patterns outside of your experience. There are usage models that exist outside of the capabilities of the larger SSDs to handle. Therefore it is cost prohibitive to have both an extreme usage model and a large capacity SSD. That's not the end of problematic environments for SSDs, either. SSDs are simply not ready, yet, for every potential usage model that exists, not for every potential environment. Therefore, you need to start asking more questions before making any further SSD recommendations.

You won't, but there it is. Now you can go back to insults and repeating exhaustive argument patterns, but I am going to leave it there. Instead of replying for another three pages all you have to do is say 'okay' and you will get the last word (as is your habit).

I admit there are users that do some scientific calculations that last 24 hours a day and push massive amounts of data which may cause durability problems with consumer grade SSDs. But those users have super computers to begin with and the price of enterprise level SSDs is not a problem for them. Most certainly, you weren't a part of this group of users according to the stats you posted yourself.

You just can't admit being wrong. Your own statistic showed you that had no problems with SSD durability in reality when the false estimated durability of 20Tb was corrected. I have no idea if it was an honest concern caused by a brain fart (you saw the amount of data you temporarily push during rendering) or if you maliciously tried to jolt your statistics by sampling only the few minutes per day that your i/o is maxed out. I can see how hard you wanted to justify your point but the fact is that for the vast majority of users SSDs are a great option. There aren't just enough hours in a day for the average person to kill the drive by writing. Even you with your atypical 10 times higher writes per day will have a 20 year estimated life on a consumer level SSD. 20 years is a darn long time when speaking of computer hardware.

Where did I insult you? Your posted stats lead to only one logical consumption - either you're jolting data deliberately i.e. cheating/lying or you did not understand that you have to monitor more than 8 minutes to find out your true i/o load. You also refused to still say where on earth you managed to find a SSD model capable of mere 20Tb writes as reflected in your durability estimate in those stats.

So you see, I have no reason to say 'okay' when I just proved your point wrong. You're the one who should say 'okay I was wrong' and walk away. You have posted zero proof, zero links - nothing to back up your arguments so far.

[BoilerDown]

Well I pulled the trigger on a single 980 when it came in stock for a short time yesterday at Newegg.  The "whether to buy" drama is over, for me.  Now comes the "how well does it actually work" drama.

[MrRiplEy[H]]

Well I pulled the trigger on a single 980 when it came in stock for a short time yesterday at Newegg.  The "whether to buy" drama is over, for me.  Now comes the "how well does it actually work" drama.

Good luck. I hope you have a faster than average computer to run it to get the full benefit. Personally I would have picked the 970 instead - only a little less performance but a lot less price.

[Gman]

The good news with these early ref 980s is the reports coming in on overclocking are excellent.  The lowest priced eVGA 980 is oc/ing up into the 1450 range and 7800 to 7900 oc/ing on the RAM very consistently.  The 970 hasn't o/c'd quite as well, the eVGA 970 superclock which is the most expensive 970 from eVGA isn't seeing the massive o/c gains that the 980 is, however as stated, for the $, it's fantastic, and due to the low price, it's going to end up SLI'd in a lot of systems for guys looking to run 4k sub 2000$ systems.

So, when you get your 980, if you overclock it, I'm interested to hear how the Precision utility works if you're going to use that, and see if you can match the reviews/reports out there of 1460 and 7800 to 7900 being easy to achieve. 

[BoilerDown]

A nice thorough overclocking review:

http://www.hardocp.com/article/2014/10/08/nvidia_geforce_gtx_980_overclocking_video_card_review/12

To sum up, the reference GTX 980 pulls away even more from the GTX 780 Ti and especially the AMD 290X when overclocked.  Its an overclocking beast, and non-reference designs should achieve far more when overclocked.

[Chalenge]

Not surprising really. I think it's the new features of the card that are the best thing about them. If you already have a 780 it simply does not make sense to upgrade, but rather wait for new drivers to include the additional features of the 980, unless power and heat are borderline issues (the 980 requiring less power and generating less heat).

Anyone looking at buying one should be aware that there are several on Amazon priced far over market value. Patience is a virtue in this case.