AMD New Horizon Event - Breaking News

[jimmysmitty]

The TB standard is better than USB speed wise, even 3.2 it seems.

ThunderBolt isn't only about speed, its other major benefit is that since it is fundamentally external PCIe with optional DisplayPort stream multiplexing, it eliminates the whole USB protocol stack. This means lower CPU utilization, lower overheads, direct DMA transfers between devices instead of strictly host-target with USB, etc.

For the regular end-user, it doesn't make much of a difference. Now that a large chunk of the cost overhead is gone though, it is that much closer to being viable in the mainstream. With both ThunderBolt and USB3(.2) using Type-C connectors, this could get confusing and possibly frustrating.

Agreed. I have been onboard with TB for a long time. Of course I still want the true TB which is fiber based but costs prohibit that.

Having both will be confusing as people wont check the cable for the TB logo to verify support but hopefully the royalty free TB makes it easier to adopt and just slowly kills off USB and replaces it as the primary connection interface.

 

[InvalidError]

but hopefully the royalty free TB makes it easier to adopt and just slowly kills off USB and replaces it as the primary connection interface.

USB isn't going anywhere: USB puts the bulk of the smarts that run the protocol on the host and its OS so endpoints can be relatively dumb, low-power and low-cost.

With ThunderBolt, all hops have to support full wire speed (unless you don't mind reduced performance starting with the first lower-speed hop) and a considerable amount of extra smarts to handle a multi-master DMA environment. That's far too much overhead for a mouse, keyboard, printer, scanner and other consumer peripherals where low cost and low power are far more critical than performance - assuming there even is any remotely meaningful performance gain to be had.

As for "true TB" on fiber, electrical TB is still 'true' TB. Same protocol, different physical layer - whatever PHY layer (micro-coax, hollow RF wave guide, twisted pairs, microstrip, fiber, etc.) cable manufacturers want to use since the PHY chips are built into cable ends, similar to how you can mix-and-match the 10+ flavors of 10GbE on switches and routers with SFP+ ports - doesn't make any of them less 10GbE.

 

[stdragon]

TB basically externalized the PCIe bus. But beyond direct attached storage and GPU breakout boxes, the most practical application is port replicator boxes for laptop docking. the USB-C connector also supplies enough wattage to charge the laptop. All this with one single connector. And, you can tap into the iGPU directly for multi-monitor support. Trust me, not having to use those USB "DisplayLink" virtual GPU adapters saves a lot of headache. They don't handle OpenGL well or at all or anything GDI accelerated.

 

[jimmysmitty]

TB basically externalized the PCIe bus. But beyond direct attached storage and GPU breakout boxes, the most practical application is port replicator boxes for laptop docking. the USB-C connector also supplies enough wattage to charge the laptop. All this with one single connector. And, you can tap into the iGPU directly for multi-monitor support. Trust me, not having to use those USB "DisplayLink" virtual GPU adapters saves a lot of headache. They don't handle OpenGL well or at all or anything GDI accelerated.

They do but man they still have a lot of issues to work out. Dell has moved their Latitude line to USB-C connections and I have had a few weird issues with them.

 

[stdragon]

[They do but man they still have a lot of issues to work out. Dell has moved their Latitude line to USB-C connections and I have had a few weird issues with them.

Depends on which USB-C adapters and laptops. I've seen many UCB-C implementations that don't support Display Port. So effectively, it's still a virtual GPU using "DisplayLink". If it's true DisplayPort pass-through, then the port replication/adapter will pass iGPU direct to the monitors.

USB-C is USB 3. Bot not all USB-C ports supports either TB and/or DisplayPort. Iconography such as the Thunderbolt or DP logo can help identify what ports are capable of what. That, or refer back to the laptop owners manual.

Truth be told, there's a lot of misconception of what USB-C can do. The answer truly depends on the make/model and which ports of the device. It's also a giant consumer headache.

 

[InvalidError]

USB-C is USB 3. Bot not all USB-C ports supports either TB and/or DisplayPort. Iconography such as the Thunderbolt or DP logo can help identify what ports are capable of what.

Type-C is nothing more than a connector spec, it has nothing to do with USB3 other than being introduced alongside the 3.1 spec as a compact and more versatile replacement for the bulky micro-B 3.0 connector. You can have devices with type-C ports that only operate at the USB1.0 speed of 1.5Mbps half-duplex using the legacy USB signal pair and nothing else. Many entry-level to mid-range phones and tablets with type-C ports only support USB 2.0 speeds and wiring.

It is ironic how USB which was supposed to make everyone's lives simpler is bloating in so many different directions that even the geekier population is losing track of its mounting pile of caveats as more alternate modes get added to it.

On the plus side, as more HSIO lanes become multi-standard, we'll get more type-C ports that can do multiple different things and hopefully alleviate the frustration of finding which ports can handle what.

 

[jimmysmitty]

[They do but man they still have a lot of issues to work out. Dell has moved their Latitude line to USB-C connections and I have had a few weird issues with them.

Depends on which USB-C adapters and laptops. I've seen many UCB-C implementations that don't support Display Port. So effectively, it's still a virtual GPU using "DisplayLink". If it's true DisplayPort pass-through, then the port replication/adapter will pass iGPU direct to the monitors.

USB-C is USB 3. Bot not all USB-C ports supports either TB and/or DisplayPort. Iconography such as the Thunderbolt or DP logo can help identify what ports are capable of what. That, or refer back to the laptop owners manual.

Truth be told, there's a lot of misconception of what USB-C can do. The answer truly depends on the make/model and which ports of the device. It's also a giant consumer headache.

We are using the Dell WD15. They do not use a virtual GPU but do use a USB NIC and audio pass through, which I don't care about. I know what you mean and also hate the virtual GPU ones that you have to install a secondary display driver to work, they work like crap.

They do also have the TB16 which is Thunderbolt based using Type-C but its also $370 bucks each.

 

[stdragon]

It is ironic how USB which was supposed to make everyone's lives simpler is bloating in so many different directions that even the geekier population is losing track of its mounting pile of caveats as more alternate modes get added to it.

USB standards aside (and they do matter, don't get me wrong), the real ball of confusion involves the Alternate Mode partners. It definitely gets out of hand when you factor in all the permutations of capabilities and their revisions PER PORT. And you're right, USB-C is just a connector. It's ubiquitous with USB 3.x standards regarding PCs and laptops, but yes, it doesn't have to be as you've indicated with cell phones for example.

Long of it short - Anytime someone asks "What all does USB-C support on my %insert device%"? To which I add "Depends, what does the manual say?" So yes, it's time everyone gets familiar with the notion that RTFM is back in vogue; if it was ever a waning concept.

 

[bloodroses]

Honestly, I don't think I'd use the code name of 'Rome' for a CPU project; considering it was an empire that crumbled. Let's just hope they don't name the next generation 'Nero'.... lol

 

[InvalidError]

Honestly, I don't think I'd use the code name of 'Rome' for a CPU project; considering it was an empire that crumbled. Let's just hope they don't name the next generation 'Nero'.... lol

Nah, just go with boats next and name it Titanic and hope it proves more unsinkable than its namesake!

 

[stdragon]

Nah, just go with boats next and name it Titanic and hope it proves more unsinkable than its namesake!

Intel was close with the Itanium in that it colloquial was called the "Itanic".

 

[jeremyj_83]

Honestly, I don't think I'd use the code name of 'Rome' for a CPU project; considering it was an empire that crumbled. Let's just hope they don't name the next generation 'Nero'.... lol

All empires have fallen at one point in history or another.

 

[InvalidError]

Intel was close with the Itanium in that it colloquial was called the "Itanic".

I know.

Just when it finally looked like Intel might be successful at driving mainstream adoption of IA64 with massive discounts on development platforms (I almost bought one), that ship got sunk almost immediately after AMD launched its x86-64 torpedoes.

 

[jimmysmitty]

Intel was close with the Itanium in that it colloquial was called the "Itanic".

I know.

Just when it finally looked like Intel might be successful at driving mainstream adoption of IA64 with massive discounts on development platforms (I almost bought one), that ship got sunk almost immediately after AMD launched its x86-64 torpedoes.

I almost wish we did move to a 64bit only design. But alas we are stuck with what we have.

 

[stdragon]

I almost wish we did move to a 64bit only design. But alas we are stuck with what we have.

If there's to be any change in a unified CPU architecture, RISK-V looks to be it. Check out the sponsors, it's insane.

https://riscv.org/membership/?action=viewlistings