I just read an article about DDR5 memory. It looks like the only solid info is that they are doubling the bandwidth and lowering the power consumption. I read that they are moving the voltage controller from the motherboard to the RAM it self. This probably looks good for the enterprise market but I have to wonder about overclocking and cost. Will still we be able to control the voltage? Will the added complexity significantly increase the cost of RAM in general? Will moving the voltage controller to the RAM signifigantly decrease the cost of motherboards? Any speculation is welcome.
I know nothing about most of this but these two things are almost definite:
Yes, either because it is more complex or simply because its new, remember when a good set of ddr4 was more than a GPU? Gonna happan again and probably worse.
I think the thing to ask is wtf will the clocks be. We’re at 3GHz ram.
Maybe we will see ram that matches CPU speeds and doesn’t have crap timing lol. Remember the timing for DDR4 when it came out? It was so bad that it preformed the same or worse than DDR3.
Also.likely to be a thing. I have heard Wendell many times complain about ddr4 not being much faster than ddr3 because the speed doubled but so did the latency.
I had a good set of DDR3 16gb 4x4 sticks @ 2000mhz cl 9 and my Buldozer 8150 CPU @ 4.7ghz 250 mhz FSB.
I didn’t realize how out of date my computer was and how much Buldozer sucked until my overclock perminantly failed a couple days ago. (Maybe some kind of board issue)… I ran them at those speeds since 2011 and was all like what do you mean the 8150 is crap and i should just buy an Intel!!! I literally had no idea what any one was complaining about. Well i think i must have hot the silicon lottry big time or something because i didn’t notice a massive delta in performance between my system and new hardware to convince me to build a new system, at least for daily task and gaming. I could even run multiple VMs with out even noticing a hit in preformence. CPU based rendering is totally different issue of course but my results were a always way better in real life than in synthetic benchmarks. I think a lot of benchmarks just see old hardware or "known bad"hardware and just spit out a low score possibly to convince you buy new stuff… Now i will build a new computer to get more cores for more virtualization and rendering / encoding and PCIE 4.0 for NVME and new GPU’s.
I don’t do anything too strenuous only some gaming and browsing really, so I have not been to bothered by the lower performance. I am using an FX-6100 @4.0GHz and 1600MHz DDR3.
I don’t go for the highest detail or anything either, using a 4GB R9 290. So over all I am happy with this just doing what it does. Though I am really looking forward to when I can upgrade to Ryzen.
I don’t know why but detail is extremely important to me. I had no problems with 4k gaming on my system when I had a Vega FE. But I think I ended up pushing my system to hard retuning it when I got the card. Or its just starting to show its age…
Recently I messed with the RAM settings and almost couldn’t recover my system. Now my system wont stay stable above 4.22Ghz CPU FSB 200mhz and 1600Mhz RAM cl 8 T1 command rate. 4k video tears and freezes now… Some games crash immediately. Some take my PC with them.
I never got that Vega FE stable even stock though now I’m thinking it was to some extent my OC and incompatibility with post Ryzen hardware. The card would halt my system at random load or not. Gaming didnt seem to make a difference in stability the card looked like it would just randomly down clock in the logs. Apparently some cards had been RMAd for that with no solution so I returned it while I still could… Perhaps the solution is to build a new PC but I would imagine that AMD cant go around just telling people to fix it by building a new PC…