Between inflation generally, and DDR4 being obsolete and unsupported by current desktop or server CPUs, it would be unsurprising for DDR4-3200 DIMMs to be at an all-time high even without the current DRAM price shock. You can never count on old memory types dropping to bargain prices, because the major manufacturers are always eager to migrate the bulk of their production capacity to current-generation memory.

That process usually takes years.

Here the price hike was pretty instant as secondary effect of DDR5 evasion in two waves. I July and now in October.

There is usually no shortage of old working PC components as they also are avalable used and tested from people decommissioning and upgrading systems. These are not some rare parts in normal market situations.

I made a habit of maxing out motherboards a year or two before upgrading to an new platform. This was always dirt cheap until like 5 years ago.

Yeah the cheapest time to buy old tech is always just when the new stuff has come out. That's when suppliers are trying to shift old stock at cheaper margins.

You can take a look at the 5800X3D and how it was at its cheapest about 2 years ago when AMD was winding down production and Zen 4 had been launched.

To further your point, a used 5800X3D still goes for ~$300 when you can get a brand new 7800X3D for the same if not slightly cheaper (was on sale at MC for $280). I assume the high cost of DDR5 is pushing more people to not upgrade to AM5 and stay on AM4 for as long as possible. And most people are avoiding Intel chips out of principle after the chip degradation debacle - you can see that based on how much lower Intel motherboards are even though they have considerably better feature sets than the AMD equivalents.

Main reason for Intel boards being cheap is because they're practically one and done because of Intel's insistence on changing sockets every other generation.

Dang, I'm going to have to put my old RAM on ebay. I thought it was worthless, but I was clearly wrong.

anyone want to buy a 2x16 GB DDR4-3200 kit that only fails memtest86 some of the time?

I wonder if semi-reliable RAM could be made to work for training. After all gradient descent already works in a stochastic environment, so maybe the noise from a few flipped bits doesn't matter too much.

Also, depends on the nature of the error. If only a small memory range is affected, you could patch the kernel to avoid it.

No need for patching you can disable specific ranges on Linux using the memmap kernel parameter. It's often used for that purpose.

Sorry but I am looking for reliability. Do you have any stick that fails 100% of the time?

Have you tried underclocking or loosening the timings?

my experience is if even memtest86 fails the memory is truly borked.

Well, it kind of depends. With XMP (which is overclocking) I've found plenty of kits on Ryzen not passing memtest with the XMP settings. Different CPUs seem to be able to run their memory controller harder without error.

And then there are other factors like more sticks of ram stressing things further. I had to downclock to get memtest stable when running 4 sticks even though each kit ran fine on it's own. But that is expected as well as 4 sticks stresses the memory controller even further.

I confess I don't have any real recent experience with DDR5 though, mostly with DDR4 on Ryzen 1000-5000 series.

Had CXMT not stop DDR4 manufacturing earlier this year, they'd at least cornered that market.

Unfortunately now they're too late on making the bank with DDR5, too: https://www.reuters.com/commentary/breakingviews/chinas-chip...

My 64 GB of DDR4 3200 RAM has tripled in price to $600. I am seriously considering selling it and just waiting for prices to drop again.

Now happy I bought that 256GB of DDR4 with the MB/processor combo of ebay last summer.