Certainly, my experience is mainly connected with metal-like ceramic materials, such as carbides, nitrides, borides etc. Although, more likely the recrystallization process and long-term creep are independent phenomena, it is possible to conclude about some correlations between them, as the mechanisms of both of them could be governing by the same moving forces directly connected with self-diffusion processes.

I am not sure that recrystallization mentioning is relevant here since a driving force for this phenomenon is a specific energy of lattice dislocations (work hardening in simple), however in creep plastic strains is rather small to accumulate high enough dislocation density.

On the other hand, grains may undergo a so called normal growth driven by a specific energy of grain boundaries. If so, the same effect should take place even without load (and hence creep) at the same temperature and time lapse.   

The creep test is held at = or  > 0.5 Tm, where the recrystallization can occur  at 0.4-0.6 Tm. The grain growth during creep testing, presumably due to secondary recrystallization. Below 0.68 Tm, grain growth lead to increased creep rate and above 0.79 Tm grain growth is more likely to occur.

The kinetics of grain growth during creep test can be noticed that by studied by statistical microstructural methods, as well as that the time of creep testing at these high temperatures is long and grain growth is more likely to occur. 

Зависит от внешних параметров (нагрузка, температура, состояние металла, внешняя среда) и типв упрочнения (легирования) металла. Существует множество данных для сталей и тугоплавких металлов (испытания длительностью до 5000...10000 часов при температурах около 0,5 Tm)

Depends on external parameters (loading, temperature, state of metal, external environment) and tipv work-hardening (alloying) of metal. There is a great number of information for staley and refractory metals (test duration to 5000...10000 h at temperatures about 0,5 Tm)

To Dr. A. S. Hammod: Please, explain what the materials you have mentioned in your message.

Although, for any materials we know the creep test can be held in very wide temperature range and the recrystallization can be observed from 0.3Tm up to Tm...

This also concerns Mr. V. V. Shirokov...

The last remark to everybody: to operate with Tm for alloys is not correct as they possess usually solidus and liquidus temperatures, but NOT melting point.

Please see Chapters 3-8 (W, Re, Ta, Mo, Nb and other refractory metals) in Volume 1 of my book on UHTM available on my ResearchGate account, and also the extremely useful link: http://engineering.dartmouth.edu/defmech/

Help the mogut cards Asby (Фрост Г.Дж., Эшби М.Ф. Карты механизмов деформации) http://mash-xxl.info/info/576784/

Dear colleague,

In principle, during a creep test devoted to obtain the parameters controlling creep, the microstructure of the sample sould remain constant, in particular the grain size. This is why the creep parameters are very often measured for different grain sizes, beacuse it is assumed that during the complete creep test the grain size must not evolve.

However, this is not always true. When measuring creep, the mechanisms for plastic deformation at high temperatures are activated, so basically diffusion of atomic species (by vacancy diffusion), as well as climbing of dislocations, which very often is superimposed to some gliding of dislocations. When these mechanisms are activated for high stresses at high temperature, or when the deformation obtained during the creep test is very high, the activation of such mechanism will produce an intensive motion of dislocation by a combination of climbing and gliding, which will be estimulated by vacancy diffusion, particularly in proximity of the grain boundaries. In such cases the creep deformation can be associated to grain growth processes and a recrystallization could take place.

If a such recrystallization takes place during your creep test, the parameters obtained from the creep test would be wrong, because as I commented the creep test must be performed at constant microstructure.

However, if such creep test corresponds to a working condition of an in-service material, this means that the mechanical properties would be probably degradated due to the uncontrolled grain growth.

I hope this could be helpful. 

It depends on metal, microstructure at start ot test, temperature, and also what is accurately meant by long term creep tests, very coarse grain size ... According to my experience of creep of low-alloy Cr-Mo steels, there is no grain growth due to the creep process for these steels. Even after 100,000 hours at 0.45 Tm. Only, the time to rupture is varying for a coarse-grained (from the beginning) sample compared to a fine-grained sample. Same for tests at 0.53 Tm for Cr-Mo steels and at 0.85 Tm for low C-steels (duration of tests in minutes, maximum 1 hour). If, in function of the temperature and the other creep conditions, some grain coarsening is observed, it is due to the temperature, not the load, as mentioned by A .A. Zisman. There also might be some grain growth on processed metal due to annealing, e.g. on ECAP (Equal Channel Angular Pressed) Aluminium alloy (6063-T835) at 473 K compared to the same coarse-grained unprocessed Al-alloy. Also, in 14MoV63 bainitic steel, the bainitic microstructure will change to a mixture of ferrite and carbids during exposition at high temperature. I note that, contrary to the above quoted metals, nanocrystalline  ceramics (TiO2) undergo "dynamic" grain growth during creep deformation.

I hope these following links could be helpful in this subjec:

1-https://books.google.com/books?isbn=1468491202

2-https://repositories.lib.utexas.edu/handle/2152/9678

Much of the confusion (at least to me) that a reader may face while following this question is that the material of specific interest is not highlighted. I have no experience with creep but have done work on superplasticity. I worked with both that are processed for static recrystallization and those that are processed for dynamic recrystallization. Under comparable testing conditions the latter have significantly higher elongation than the former evidently due to repeated cycling of recrystallization-substructure growth-recrystallization process that makes the material sustain much greater strains. I have uploaded a paper on Researchgate that highlights this issue. In conclusion I would like to state that there is no unique theory that can explain grain growth (or lack of it) for ALL materials. The answer should be found for the particular material undergoing creep. 

Many thanks to Dr. A. S. Hammod for useful links.