It is well known that the Large Cardinal Axiom which asserts the existence of a measurable cardinal number is inconsistent with V=L. Do any of the wellknown Large Cardinal Axioms contradict V=OD?
Almost all of the usual large cardinal axioms can be preserved by the reverseEaston support class forcing to make every set of ordinals coded into the GCH pattern, or the $\Diamond^*$ pattern, and these axioms imply V=HOD.
For example, this is true of the supercompact cardinals by the usual Laver argument, and analogues work with many other large cardinals, including much larger cardinals. Therefore, any smaller cardinal notion that is a consequence of supercompact cardinal or these other large cardinals will also be relatively consistent with V=HOD.
Meanwhile, there are some large cardinal notions that imply V=HOD outright. For example:
Theorem. If there are a proper class of Laverindestructible supercompact cardinals, then V=HOD.
Proof. If $\kappa$ is supercompact and Laver indestructible, then for any bounded subset $x\subset\delta<\kappa$, there is $<\kappa$directed closed forcing $\mathbb{Q}$ that codes $x$ into the GCH pattern above $\kappa$. If $\kappa$ is still supercompact after this forcing, then it would be $\Sigma_2$reflecting in the extension, and the existence of a coding place for $x$ would reflect below $\kappa$ into $V_\kappa$, which was not changed by the forcing. Thus, $x$ is already coded in $V$, and hence every bounded subset of $\kappa$ is coded into the GCH pattern below $\kappa$. Since we have a proper class of such $\kappa$, it follows that every set is coded, and thus V=HOD. QED
It also works with indestructible strong cardinals and many others; really all you need is that the $\Sigma_2$correctness is indestructible.

$\begingroup$ The theorem is false with only a bounded number of indestructible supercompact cardinals, since you can add a Cohen set above them all, and this preserves indestructibility while killing V=HOD. $\endgroup$ – Joel David Hamkins Aug 11 '16 at 17:08

1$\begingroup$ A similar argument shows that the Shelah philosophical attitude toward the GCH  that it holds in a "random" pattern with all possible behaviors occuring somewhere  implies V=HOD. $\endgroup$ – Joel David Hamkins Aug 11 '16 at 17:22

$\begingroup$ Many thanks for this detailed response. It was a surprise for me, since I had expected that most of the Large Cardinal axioms would contradict V=OD outright. It shows how little I know about that area of Set Theory $\endgroup$ – Garabed Gulbenkian Aug 12 '16 at 18:56

$\begingroup$ These Large Cardinal Axioms bring many more sets into existence and I thought that quite a few of them would not be Ordinal Definable. $\endgroup$ – Garabed Gulbenkian Aug 12 '16 at 19:00