Recent content by victhor393

I decided to bite the bullet and soldered a couple of jumper wires directly to the board. I had no wirewrapping wire or magnet wire around, so getting the wires soldered to that tiny header was a small miracle, let alone that there were no shorts, on the first try, no less: I plugged the card...

My programmers have arrived today, however, I haven't got the adapter PCBs to plug into my board yet. I can't test them right now, either way, as I'm a little too busy. I'd have to consider if it's worth it to solder jumpers directly to the board... or if I should get the adapter PCBs made...

I just bought a Xilinx clone programmer, a CH341 programmer, and a SWD programmer. Now I can finally take this project seriously!

@mwrnd Great work. I need to find some time to work on this project. It's been years since I bought this board and did absolutely nothing with it... @oneplane Yes, there is a microcontroller on the back of the board. It's an Atmel SAMD20, named U15 on the board.

I tried to trace the pins for the top JTAG port, but I was unable to. There's no continuity. It seems there's something between the Zynq and the port, but I don't know what it might be. I found these pins by overlaying the pin diagram available on one of the Zynq Ultrascale datasheets over a...

If he has the time and interest to look into this, why not? Seems like he's done a lot of work figuring out those Mellanox cards.

I have a Broadcom BCM957404A4, a Chelsio T520-SO-CR and a system with claimed ASPM support, sadly, no way to measure power consumption. BCM97404A4 reports ASPM is supported and enabled, T520-SO-CR reports ASPM is not supported. This is the output from lspci -vvv for the cards, the link is...

They won't, they are for AMD coolers and not for LGA2011 coolers.

I can vouch for these adapters as well, however I used tower coolers (Hyper T200) with mine, unmodified. These maintained the E5-2670v2 (115W) at around 40C above ambient or thereabouts...

Congrats. These are too interesting to go to waste.

I want to say that it should work, but I can't say that for sure. A higher speed module should still work, it may even have SPD timings for 2400 MHz. I don't have experience with this configuration, only with mixing speeds (2666 with 2400) on a different platform and it does work that way, dual...

As an interesting historical note, some of the T3 generation cards had a FPGA on board. It replaced the discrete TCAM that used to be present in full-featured T3 cards. The FPGA interface can probably be used without the need for SERDES, so that's convenient, since it potentially means lower...

Maybe that product briefing is missing some detail. As always, documentation held under confidentiality agreements rears its ugly head over those trying to better understand what goes on inside the magical black boxes... Note how it has no mention of: buffer size, TCAM size, fabric bandwidth...

U.2 have a finned enclosure sometimes, but you're right. They are designed to be cooled by forced air. Actually if you are dead set on buying U.2s, the reason why these enterprise drives have such weird capacities (960GB, 1.92TB so on) is that they have a little (or a lot more, for "write...

I honestly think your fear of TLC may be misguided, high end TLC drives are plenty fast these days. MLC is seldom used anymore, outside of enterprise drives (though even these are turning to TLC, thanks to improvements in performance and a growing need for density) and embedded devices. But...