ir_cow

Ive made the same mistake plenty of times

Well it should go at 100 MB/s or higher. Do you have the chipset drivers installed? I know it sounds stupid but if you don't have the right drivers than it reverts to USB 2.0 and uses Windows defaults ones. The other option is one drive is bad, which is why the read (or write) speed is so low. I guess it also could be if you are copying lots of little files than it will give you a false number. Try copying one big file and see if goes. Lots of choices....

u Crazy!! Hows the HDR working for you? I have issues with games saying its in HDR mode but be completely blown out until I go into the windows 10 HDR menu turn it off and than back on. Works for a few days and the same thing happens. So now I leave HDR disabled. Not worth the hassle of having stuff looking wacky half the time.

If your computer shuts down browsing facebook it may be a bigger problem. Have you done any stability tests? Anyways something comparable to a ATI Radeon HD 4550 512 MB would be.... I honestly don't know on the ultra low end. What country are you buying this in?

Well thats awesome EK paid the bill. Still seems a bit overpriced, but thats EK for ya lol. Unless you saying the whole build was $1700 CAD. Thats a different story.

Yeah watercooling isn't what it use to be. Now its more of a novelty at this point. It diffidently can give a much cleaner build, but as far as performance goes the gains are minimal compared to the cost. None of these Ryzen 3000 chips overclock well. You might be able to nudge another 100mhz over AIO, but once you pump 1.4v into the CPU for 4.4Ghz it will get toasty. Edit: I hope you didn't pay $1700 for the watercooling part. If you went all out and did the motherboard too, Its like $600

No you didn't do anything wrong. Every card is different so you will have to do more stability tests to really get the most of it out. Usually artifacts without actually crashing is memory errors. I use different methods for overclocking because the same software it doesn't work on every card for some reason. 1: Download / MSI Afterburner (or Precision X1). Run it and set Power and Temperature limit to the max. Leave everything else alone. 2: Open MSI Kombuster and set it to Artifact Scanner mode. Than run stress test. 3: If you overclock far enough on memory you should see a error count. (I will upload a screen shot). If so than this is only program you need. Otherwise it is really just a tedious process and kinda done blindly becuase Kombuster is a vary handy program, it just doesn't always work with certain cards. The next step is straight forward. Raise the memory until you get errors, than back it off. Now set the memory back to stock and do the same thing for the GPU clock. Usually it will just crash instead of error. So find the maximum of that value. Now step both overclock values and see if you get errors. Most of the time you will see errors, so you will either need to back off the memory or the core clock. Usually the core clock is the one that needs to be stepped back 10-15 Mhz. 4: If Kombuster doesn't work, you will just have to blindly set values, run Kombuster is normal stress mode and when it crashes, just back it off a bit. Once you find you found good values, run 3DMark Time Spy Extreme stress test. It will do 20 loops. If it passes with 98% or you are good. 99%+ is perfect. Anything below 98% I would say will yield lower FPS because of the extra errors introduced. This is my 2070 SUPER FE settings. I had it at 155 Core Clock, but it gave a 97% in the stress test. 145 gave me 99.1%. Upping the voltage was actually the worst offender, so more voltage isn't always the answer. NVIDIA RTX 2070 SUPER 145+ - Clock 650+ - Memory Stock - Voltage 121+ - Power Limit (Max the card allows)

Well currently the i9 9900K is still the fastest gaming CPU (better than any Ryzen) if that helps you make a decision. Im not sure how the i5 6600k stands up against AMD 3700X or i7 9700k which is fairly comparable in games. I would have to do some research. I dont have that cpu to do a direct comparison.

I've been eyeballing this Patriot Viper Steel 16GB (2 x 8GB) 288-Pin DDR4 SDRAM DDR4 4400 (PC4 35200) memory kit for about 2 months. I want to replace my test system memory. It has slowly gone down in price. For the Ryzen Platform it is overkill but It can easily be tweaked to 3800 CL14. The only retail kit that can do this is the G.SKill NEO 3800 and costs about $100 more. If you don't care about RGB lights, this is a steal. I'm sure other kits can do DDR4 3800 CL14, but they all cost more. Best I can do with my G.SKill Ripjaw V 3200 is 3800 CL16 @ 1.42v or 4266 CL19 and that kit still costs more! I'll find out in a few days how plug and play these are. But considering its high binned Samsung-B die, I cannot see this being a problem. It might even boot to 4400 on the ASUS 570X ROG Crosshair Hero. We will see... Edit: For those gamers, from the tests i've ran, 3000 to 3800 has little impact on FPS. Mostly within margin of error. Above 3800 the FLCK ratio breaks and a latency penalty is introduced. I haven't done much testing on this part. I'm hoping to get a article out by the end of the year. Nothing ground breaking as other people have already done similar tests.

If its LED strips than it will plug into the 5v or 12v RGB headers. For fans it depends on if its true RGB fans or just ones that plug into the fan header. You have a link to the fans you bought?

I did a crappy article a few years ago when I compared a i7 3930K to a i7 7700K. The idea was solid, execution was flawed as I used a GTX 1070 and didn't compare enough games. Anyways while the CPUs had similar benchmark results. Old 6-core vs new 4-core. The big difference was switching to DDR4. I saw 30% increase in FPS and the limit was actually the video card I was using didn't reflex how much further it "could have been". We all know taking a RTX 2080 Ti and pairing it with a i5 3570K would be counter-productive. But a middle ground does exist. No need to buy a whole new system. Just replace the CPU+MB with a Ryzen 3600 or a i5-9600K.

You picked a good brand, I guess It depends on if you want the RGB. Price wise, it seems all the RGB lights start around that price of $80. Newegg has (none-RGB) G.Skill 3200 CL16 for $70 and 3600 CL16 for $90 ($100 with RGB). As for performance. When I benchmarked 3800 CL16 ($200) vs 3200 CL14 ($150). It was a big difference is synthetic benchmarks, but in real-world ones (aka games), 1920x1080 resolution saw 2FPS increase with a RTX 2080 SUPER in certain games like The Division 2, other, no difference. Considering both kits are double the price of a 3000 kit, the price isn't justifiable unless overall the build cost is higher. Edit: I wanted to see what DDR4 3000 CL16 would be like so I tested it in Borderlands 3 with a RTX 2070 SUPER. Settings: 3733 CL16-16-16-35-50-T1 / FCLK 1866 3000 CL16-16-16-35-50-T1 / FLCK 1500 AVG FPS. Graphical: Badass 1080P 3733 - 84.35 3000 - 85.36 2K 3733 - 60.48 3000 - 61.73 4K 3733 - 33.24 3000 - 33.98 All three within margin of error. The faster memory is done after hours of benchmarking already and card is warmed up. The other I just did with a cold video card. That might be why the frame-rate is slightly higher. So the answer to the age old question of if memory speed matters. It for the most part does not in PC gaming. This is only one game I quickly tested and of one video card, but from the other tests I have done, the difference can be chalked up to margin of error. I'm sure it has a higher impact on lower tier video cards and the Ryzen 3600, but for the 3700X and above, it doesn't matter.

I never had one so I cannot say what I don't have in it . From the stuff I've read and past NZXT cases I have reviewed, this one is fairly good for the price. Personally I think it looks really nice. One of my most favorite cases in the Phanteks Luxe. I think it's been replaced with the Luxe 2. Not super cheaper, but really good airflow and some basic lighting. Do you have a certain price point or just looking for the best air / performance?

I just used the newest version of ATiFlash today. It works with windows 10, maybe past versions didn't but this current one sure does. You also don't have to use DOS to force flash as people think. just run the windows program in the CMD. Buyback just means the customer traded the old card for discount while buying a new one. Usually it means they had to buy it from the shop in the first place. Anyways if the shop owner says they aren't used of mining, he wouldn't (or couldn't) know unless the previous person didn't flash the BIOS back. I don't think it would be a issue since the owner says it isn't used that way and hopefully backs it up with a warranty. In either case, if the shop has a warranty than your fine. Really no matter what card you buy used, it is impossible to know how it was used before.

Hmm... Well I really liked the NZXT H500. Its now been replaced by the H510 which I know nothing about. Phanteks is always a good choice because of the build quality. I know nothing about the two Thermaltake cases.