AMD's new EXPO Ultra Low Latency memory promises tighter timings without manual tuning. We tested G.Skill's latest DDR5 kits to see how much faster they make Ryzen gaming.
Testing AMD's EXPO Ultra Low Latency Memory
- Thread starter Steve
- Start date
This just confirmed that I don't need to do any manual tuning for my CL30 32GB RAM. thank god I pulled the trigger right before the prices exploded.
OortCloud
Posts: 2,264 +3,277
Here's some more things the AI all-hype-no-substance price fixing cartel have stopped you from having...
maxxcool7421
Posts: 892 +1,294
Keep in mind this tech is not for 'us' the enthusiasts. This is about APU/APU-AI optimization where 5% across 10,000 nodes makes a 6 figure profit difference to an AI farm.
onebignerd
Posts: 154 +89
Vanderlinde
Posts: 807 +532
Before people start jacking up their timings, properly test these before actually applying. Tools like memtest86 on a stick is what you need. A good 3 solid passes or better 24 hour pass means your timings are extremely stable.
If it fails within 5 minutes of testing your pretty much at the edge already.
I could not afford an unstable system causing all sorts of stuff.
If it fails within 5 minutes of testing your pretty much at the edge already.
I could not afford an unstable system causing all sorts of stuff.
Call me a Nazi if you like, but omg, I hate the misuse of the term "validation".
Validation is the process of determining if a requirement is "valid" and applicable based on its application, or particular "use case".
Verification is whether or not that requirement actually meets the required specification (through testing).
This drives me up the freaking wall, with nearly every engineer and "architect" in the companies I've worked for consistently get it wrong. The words are so close, you'd figure people would get it right HALF the time, but they don't.
Validation is the process of determining if a requirement is "valid" and applicable based on its application, or particular "use case".
Verification is whether or not that requirement actually meets the required specification (through testing).
This drives me up the freaking wall, with nearly every engineer and "architect" in the companies I've worked for consistently get it wrong. The words are so close, you'd figure people would get it right HALF the time, but they don't.
Endymio
Posts: 5,288 +4,901
Circular definition much?
Being also myself a left-brained word fetishist, I empathize, but, while to a QM engineer, verification and validation are distinguished separately, but in common English parlance -- such as this article -- the distinction, if any, is sterile.
mdenigma
Posts: 44 +44
To sum it up: The X3D-Variants are the best choice for gaming. Outperforming in every use case over the heavily memory-optimized Non-3D-Variants. Memory-Tuning on X3D usually leads to measurable, but not significant performance uplifts.
My Recommendation: Do not spend a huge amount of money or time into memory-tuning that. X3D performs decent with CL40 and all your games will run just fine and stable (!). EXPO increases voltage way above 1.10V and is not worth the performance-bump.
My Recommendation: Do not spend a huge amount of money or time into memory-tuning that. X3D performs decent with CL40 and all your games will run just fine and stable (!). EXPO increases voltage way above 1.10V and is not worth the performance-bump.
I believe the author knows very we the distinction between the two words based on how the article is written. Validation is just the correct word that has always been used in the lexicon of technical jargon. We. There are lots of examples of terms used in this way throughout the Tech community at large. I mean we all know a bug in the system doesn't mean you are going to find a fly in your case although that is how the term originated. Validation is and always has basically meant verification in tech talk and in most cases the two could certainly be used interchangeablly. Both would be right.
Similar threads
- Locked
