I just found out about BLUESCsi and wow, I didn't know it.
I was familiar with SCSI2SD, which has now reached version 6. However, I always find myself surprised and a little disappointed by these technologies. SCSI was a beautiful and robust technology that reached peak speeds of up to 320 MB/sec.! Yet, these modern SCSI-2-SD products that are being developed persisting in not breaking the 10 MB/s barrier.
I do not understand why. Even previous versions of SCSI2SD stop at 1 or 2 MB/s!!!
I understand that in many of the systems for which they are intended, higher speeds would not even be usable by the respective processors, BUS, RAM, etc. I also understand that 10 MB/sec for vintage systems, with solid state memories at very short access times, such as SD cards, are already fine.
But I wonder: given that somebody took the trouble to designing these cards from scratch anyway, given that we are in 2023 and modern technology exists, we might as well design such adapters being compatible with higher SCSI standards and capable of using their higher speeds. I'm not saying exactly with the U160 and U320 SCSIs, but at least with the 40 and 80 MB/sec Ultra Wide SCSI 2 protocols.
Even if the oldest retro systems cannot fill that bandwidth, they would still be well usable by many vintage systems on both the PC and Mac side (fast 486s and Pentiums for example), and combined with the fast solid state memories, they would provide dazzling performance! Why limit yourself to 10 MB/sec.!??
I don't think this would lead to a significant increase in production and market costs. In fact, here we are not talking about moving, for example, today, from a 12 or even 7 micron CPU production process to one of 0.5 micron, which fundamentally does not exist yet and would require very expensive investments in research and development.
We are talking about a technology, SCSI, that is 30-40 years old. We are talking about this in 2023, an era in which we talk about hundreds of Gigabytes per second like they were peanuts (think of modern GPU memory bandwidths for example). In such an era, with such technology, differentiating an adapter by producing it at 40 MB/sec. rather than 10 MB/sec. is like changing the production of a modern car so that it has 101 horsepower instead of 100.
It's irrelevant, for today's technology.
So it must also be for development and production costs. However, it is not at all irrelevant, in fact it is fantastic, for those old retro systems that would then benefit from it. But no. They insist on stopping at 10MB/sec., in an age in which a 500 GB NVMe Gen 4 SSD that gives you 7000 MEGABYTES per second costs little more than 50 dollars. What are we talking about? Simply, why??
Do any of you have a background in computer engineering? Can any of you answer my question?
