Opteron clock throtting

Does anyone know if the Opteron (and/or the Athlon 64) has support for clock-throttling similar to what the P4 has? Another way to put my question: If I accidentally knock the HSF off the Opteron while it is running, will the processor A) slow itself down and thus reduce its temperature while still running, B) shut itself off and effectively lock-up/crash the computer, C) fry itself.

A side question is, does the Opteron (or Athlon 64) require the motherboard to do anything special in the event of an overheat (such as monitor the temperature like the 32bit Athlons did)?

I doubt they have this technology integrate to their upcomming opteron or A64 CPUs. This technology is for intel cpu only.

However, we will never know until we see the review. I think they still have to use the motherboard sensor to determine the shutdown temp.


Computer Maniac ;)

Yes it does. Some of the early models were even showcased at trade shows with the HSF completely removed - they kept working.

Links? Articles?

Quote:

---

Originally posted by ComputerManiacs
I doubt they have this technology integrate to their upcomming opteron or A64 CPUs. This technology is for intel cpu only.

However, we will never know until we see the review. I think they still have to use the motherboard sensor to determine the shutdown temp.


Computer Maniac ;)

---

No, the Athlon XP has thermal throttling as well. So I'm sure the Opteron will.

Even with thermal throttling, knocking off the hsf reduces cooling so much it may not save the cpu.

Quote:

---

Originally posted by Racer^
Links? Articles?

---

Google:

http://www.aceshardware.com/forum?read=105032449

Here's a good article in PDF format on Hyper Transport (A64 & Opteron chips) and what it does, specifically on page 6 it states that Hardware and Software Thermal Clock throttling are enabled.

http://www.hypertransport.org/docs/ht-powermgt.pdf

Google is a tool...use it ;)

google? I did use it, and the links were like the ones you posted; either second hand conjecture (which I am capable of) or marketing fluff. I was hoping for more but I guess I'll have to wait like everyone else.

P.S. I've heard of pc repair techs testing P4s that come in by removing the HSF while the computer is running. I don't know what they were testing, but it sounds like they trusted the P4s throttling alot.

Quote:

---

Originally posted by Racer^
google? I did use it, and the links were like the ones you posted; either second hand conjecture (which I am capable of) or marketing fluff. I was hoping for more but I guess I'll have to wait like everyone else.

.....

---

Hmmm what do you call a paper written by a member of AMD's technical staff?? Did you take the time to scroll down to the table on page 6?!?https://www.sharkyforums.com/

I reread the document and it still looks to me like a generalized HyperTransport marketing presentation designed to summarize the possible benefits that HT could have if implemented. Just because a feature is available in a given technology doesn't mean it is implemented or enabled. Also, I saw no mention of any examples in which the power management features listed in the article were (to be) utilized by Hammer, Opteron, or anything else of note.

Quote:

---

Originally posted by Racer^
Also, I saw no mention of any examples in which the power management features listed in the article were (to be) utilized by Hammer, Opteron, or anything else of note.

---

Did you happen to note the date of the presentation?? Or do you know if/when AMD chips started using "hyper-transport"?

Quote:

---

Originally posted by Almost Famous
http://www.imaginetstudio.com/images/todo.gif

---

hehe.

Quote:

---

Originally posted by Racer^
I reread the document and it still looks to me like a generalized HyperTransport marketing presentation designed to summarize the possible benefits that HT could have if implemented. Just because a feature is available in a given technology doesn't mean it is implemented or enabled. Also, I saw no mention of any examples in which the power management features listed in the article were (to be) utilized by Hammer, Opteron, or anything else of note.

---

The chart clearly states that Clock throttling is "hardware enforced" , as well as software enforced which suggests to me that it's in the hardware ....

Quote:

---

Originally posted by russ_watters
Did you happen to note the date of the presentation?? Or do you know if/when AMD chips started using "hyper-transport"?

---

Opteron;)

Quote:

---

Originally posted by Almost Famous

DBL Post

---

Quote:

---

Originally posted by Racer^
I reread the document and it still looks to me like a generalized HyperTransport marketing presentation designed to summarize the possible benefits that HT could have if implemented. Just because a feature is available in a given technology doesn't mean it is implemented or enabled. Also, I saw no mention of any examples in which the power management features listed in the article were (to be) utilized by Hammer, Opteron, or anything else of note.

---

And here you go straight out of the Opteron white paper here:
http://www.amd.com/us-en/assets/cont...docs/23932.pdf

"The AMD Opteron Processor provides a hardware-enforced thermal protection mechanism. When the processors temperature exceeds a specified temperature, the processor is designed to protect itself from over temperature conditions by stopping it's internal clocks and asserting the THERMALTRIP_L output."

Is that proof enough? The whitepaper on the proc? I doubt if you saw it with your own eyes you'd believe it, but there it is for ya ;)

edit: that quote was taken from 3.6 on the whitepaper.